Backlog F5: Qwen3-Embedding-4B measured, on llama.cpp and vLLM, and against bge-m3

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

.gitea/workflows/ci.yml

@@ -39,6 +39,14 @@ jobs:
       - name: Build
         run: go build ./...
 
+      # llmgw-kb runs on people's own machines, Windows included, which no
+      # test here runs on: building it for each catches a Unix-only call.
+      - name: Build llmgw-kb for every system
+        run: |
+          for target in windows/amd64 windows/arm64 darwin/amd64 darwin/arm64 linux/amd64 linux/arm64; do
+            CGO_ENABLED=0 GOOS="${target%/*}" GOARCH="${target#*/}" go build -o /dev/null ./cmd/llmgw-kb
+          done
+
       # The Terraform provider is a module of its own, which ./... above does
       # not reach. Its tests check its client against docs/admin-api.openapi.json,
       # so a change to the admin API that breaks the provider fails here. The

.gitea/workflows/release.yml

@@ -55,6 +55,21 @@ jobs:
             (cd dist && sha256sum "${name}" >> checksums.txt)
           done
 
+          # llmgw-kb, which people run on their own machines to put their
+          # documents into a knowledge base: Windows too, which llmgw itself
+          # does not build for.
+          for target in windows/amd64 windows/arm64 darwin/amd64 darwin/arm64 linux/amd64 linux/arm64; do
+            os="${target%/*}"; arch="${target#*/}"
+            ext=""; [ "$os" = windows ] && ext=".exe"
+            name="llmgw-kb-${version}-${os}-${arch}${ext}"
+            echo "building $name"
+            CGO_ENABLED=0 GOOS="$os" GOARCH="$arch" go build \
+              -trimpath \
+              -ldflags "-s -w -X gitea.mynodes.xyz/ilker/llmgw/internal/version.Version=${version}" \
+              -o "dist/${name}" ./cmd/llmgw-kb
+            (cd dist && sha256sum "${name}" >> checksums.txt)
+          done
+
           # The reverse proxy configurations, so a release page carries
           # everything install.md asks for and the repository is not needed.
           for f in deploy/proxy/Caddyfile deploy/proxy/nginx.conf; do
@@ -91,6 +106,7 @@ jobs:
           # test the cross-compiled ones, which is why the native build is
           # checked rather than assumed.
           ./dist/llmgw-${version}-linux-amd64 --version
+          ./dist/llmgw-kb-${version}-linux-amd64 --version
           cat dist/checksums.txt
 
       # Signed, so a machine that trusts the key built into it installs

BACKLOG.md

@@ -569,9 +569,12 @@ portal is not done — the page shows; the console and ctl change.
     client tests against a fake admin API and the OpenAPI document;
     acceptance tests with Terraform and OpenTofu against the fake; one
     apply, empty plan, in-place change and destroy against a real `llmgw
-    server` run by hand. Left: organizations, teams, projects, people and
-    keys are not in the admin API (they are `llmgw org`/`users`/`keys`
-    against the store), so not in the provider; a model's definition beyond
+    server` run by hand. Organizations, teams, projects, keys and
+    memberships are in the admin API since 2026-10-08 (`/api/v1/orgs`,
+    `teams`, `projects`, `keys`, `users/{id}/memberships`) and in the
+    provider (`llmgw_org`, `llmgw_team`, `llmgw_project`, `llmgw_key`,
+    `llmgw_membership`); accounts are still made by sign-up, invitation or
+    OIDC/SCIM. Left: a model's definition beyond
     scaling, class, owner and adapters is changed by replacing it (the API
     has no edit, and `resources` is not used yet); `canary`, `slo`,
     `scaling.schedule` and `prewarm`, `guard` role and `origin` are not in
@@ -793,10 +796,19 @@ leader 30s later with /health answering throughout; rollback did the same.
   facility state from the store (it kept what it read at start).
 - Retires `appendGuardIfMissing` (server init): the import does it.
 
-### E2. Generated files from settings (L) — firewall on/off done 2026-10-06
+### E2. Generated files from settings (L) — firewall on/off done 2026-10-06, every machine 2026-10-09
 `firewall.enabled` in the cluster's settings reaches nodes over the agent
-stream and control planes through llmgw-firewall-sync.path. Left: the mirror
-machine and store members, the allowed set from membership, the other files.
+stream and control planes through llmgw-firewall-sync.path. Since
+2026-10-09 the mirror machine too (the leader PUTs /v1/firewall, the mirror
+writes `.firewall.json`, the same path unit applies it) and store members
+that run no llmgw (the leader keeps it in `cluster_state`, a root timer
+reads the member's own copy with psql); `firewall_networks` reach all of
+them, and the mirror lets in the network of every node and control plane
+the leader knows. Tried on the PoC: off and on, and a network added, on
+control planes, mirror, store member and nodes. Left: the allowed set
+replaced by membership rather than added to (a network taken out of the
+settings stays let in), the other files (the mirror's unit flags,
+keepalived, etcd and Patroni configuration).
 
 Every machine runs the agent (root) with roles — node, control plane, mirror,
 store member — and the agent renders and applies what was written once by
@@ -807,17 +819,288 @@ Patroni configuration. The firewall's allowed set is the cluster's membership
 removed changes every machine's rules. Retires `firewall.Admit` and
 `firewall_networks` in server.yaml (imported into the store).
 
-### E3. agent.yaml down to bootstrap (M)
+### E3. agent.yaml down to bootstrap (M) — done 2026-10-08 (branch `settings-e`)
 agent.yaml keeps the control planes' addresses, its certificate and its id.
 Labels, fabrics, engines, heartbeat and supervisor settings move to the
 store per node and reach the agent over its stream; a change to the file is
 noticed and the agent restarts itself.
+Built as a section of the cluster's settings rather than a store of its own,
+so history, audit, `ctl config`, the console and the portal page cover it:
+`agents.default` and `agents.nodes.<id>` (config.AgentSettings; bootstrap
+keys node_id/server(s)/hostname/mgmt_addr/tls refused; each node's merge
+loaded as an agent would, so a bad fabric is refused on save). A live key:
+the control plane sends ControlMessage.AgentSettings on register and, a
+second apart per node, on change; the agent compares with what it started on,
+checks it loads beside its agent.yaml (else logs and keeps running), keeps it
+in /var/lib/llmgw-agent/agent-settings.yaml and restarts (engines adopted).
+At start: agent.yaml < kept settings < LLMGW_ env; kept settings that no
+longer load are set aside, never a reason not to start. agent.yaml is
+watched every 10 s and restarted on when it changes and loads. Left: file
+keys are not imported into the settings (they keep working under them);
+`node add` still writes a full agent.yaml; labels have two homes now
+(`ctl label`, which wins, and agents.*.labels).
 
-### E4. Web settings page (S)
+### E4. Web settings page (S) — done 2026-10-08 (branch `settings-e`)
 The same document, edited on /admin/settings, for whoever has no terminal.
+Built: administrators only (staff roles get 404: webhook secrets), checked by
+the same function as PUT /api/v1/settings, made from a base version or
+refused on a newer one (the edit kept in the editor), history with view and
+roll back as a new version, audit with the size not the body, English and
+Turkish, a Settings link in the menu.
 
 ---
 
+## F. Company documents: knowledge bases (RAG), then adapters (2026-10-08)
+
+A company wants a model to answer from its own documents. Retrieval is the
+tool for *knowledge* (sources shown, re-indexed in minutes, access per
+tenant); LoRA training is for *behaviour* (vocabulary, format) and has no
+access control — whatever goes into weights reaches everyone who may use the
+adapter. RAG first. Decisions taken: access per collection (no per-document
+ACL yet); a personal key or session names a project or team and gets that
+chain only, never the union of its person's memberships; PostgreSQL store
+only, off on SQLite with a sentence saying why.
+
+### F1. Store: collections, grants, chunks (M) — done 2026-10-08 (branch `rag`)
+Built (`internal/store/knowledge.go`, migration 0035 PostgreSQL only):
+- `kb_collections` (owner scope/id, org, embed_model, dims), `kb_grants`
+  (read|write, any scope, never across organizations; a key needs a
+  project; a person needs a membership in the organization),
+  `kb_documents` (digest: the same again is no change, a new one replaces
+  in one transaction).
+- `kb_chunks` is created when the store opens, not in a migration (a
+  migration needing an extension would stop the control plane; one that
+  skipped it would never run again): partitioned by organization (hashed
+  partition name), `tsv` generated (`simple`), GIN on the parent, one
+  partial HNSW index per collection on its organization's partition
+  (`embedding::vector(dims)`, `WHERE collection = '<id>'`). EXPLAIN on
+  20k rows: partition pruned, that index used, policy applied.
+  Later (scale, 2026-10-08): a partial GIN per collection too
+  (`kb_tsv_<id>`, backfilled at open); `vec_type` per collection, `halfvec`
+  on pgvector ≥ 0.7 (up to 4000 dims); `SET LOCAL hnsw.ef_search` to
+  candidates(k) clamped 40..1000 (EXPLAIN: 40 rows without it, all with
+  it); the text arm takes candidates by `ts_rank` in a subquery and scores
+  only those.
+- RLS FORCEd on `kb_chunks` against `llmgw.kb_org`/`llmgw.kb_collections`
+  (set with set_config per transaction), one role — the table owner — not
+  a separate one. Chunks are deleted by the foreign-key cascade from their
+  document, which RLS does not block.
+- `CollectionsFor(Effective, write)`: owner or grant anywhere on the chain;
+  no organization (personal key without a project) reads nothing.
+- `SearchChunks`: per collection nearest-neighbour (literal id and cast so
+  the partial index is used) plus a full-text arm with any word
+  (`a | b`, lowercased by the database, not Go), fused by reciprocal rank.
+- Deleting a project/team/org that owns collections is refused like one
+  with members (changed from "deleted with them": every other delete in the
+  tree refuses); DeleteUser/ForgetUser delete the person's collections;
+  deleting an organization drops its partition; grants to a deleted scope go.
+- `server init` installs `postgresql-N-pgvector` / `pgvector_N` and enables
+  it; pgha installs it on every member at setup and enables it on the
+  leader on every run. Neither stops on failure.
+Left: a cluster set up before (the PoC) gets no package from a rerun of
+pgha on a running cluster — install by hand or add an upgrade step;
+`server restore`/`CopyTo` do not carry kb tables (SQLite → PostgreSQL only,
+and SQLite has none); tests need `LLMGW_TEST_POSTGRES` pointing at a
+pgvector database as a non-superuser (a superuser bypasses RLS).
+
+### F2. Ingest on the user's machine (L) — done 2026-10-08 (branch `rag`)
+Uploading whole files to the control plane was dropped: a PDF is 20–100×
+its text, and the work belongs where the files are. Instead a small client,
+`llmgw-kb`, run by the person on their own machine with their own key:
+- **Its own binary** (`cmd/llmgw-kb`), for Windows, macOS and Linux on amd64
+  and arm64. `llmgw` itself does not build on Windows (agent syscalls), and
+  employees' laptops should not carry the server, agent and setup screens. A
+  test keeps server packages out of its imports.
+- `llmgw-kb sync <collection> <dir|file…> [--include/--exclude globs]
+  [--dry-run] [--delete]`: scan → local catalog (path, size, mtime, sha256;
+  cache under os.UserCacheDir, digests not recomputed for unchanged files)
+  → the collection's manifest from the gateway (name → digest only) → diff
+  → for new/changed files only: extract, chunk, embed through the gateway's
+  `/v1/embeddings` with the person's key (its accounting, limits, guard and
+  band apply), and send name, digest, chunk texts and vectors; `--delete`
+  removes documents whose files are gone. `--dry-run` lists what would leave
+  the machine and how much. A file that cannot be read (locked, unsupported,
+  no text) is skipped with the reason; the run goes on.
+- **Formats, as many as possible.** Built in, pure Go: PDF (text layer),
+  DOCX/XLSX/PPTX, ODT/ODS/ODP, RTF, EPUB, HTML/XHTML, EML (text and HTML
+  parts), Jupyter notebooks, subtitles (SRT/VTT), CSV/TSV, JSON/YAML/XML,
+  Markdown/RST/AsciiDoc/Org, plain text and source code; ZIP archives
+  recursed into. With a helper found on the machine: legacy DOC/XLS/PPT, RTF
+  and anything else LibreOffice opens (`soffice --headless --convert-to`),
+  images and scanned PDFs through Tesseract (`tur+eng`). Text in UTF-8,
+  UTF-16 (BOM) or Windows-1254/1252 is converted to UTF-8.
+- **Across systems:** document names relative to the synced root, `/`
+  separated, Unicode NFC (macOS hands out NFD), a warning for two names that
+  differ only in case; CRLF → LF; key from `LLMGW_API_KEY` or `llmgw-kb
+  login` (os.UserConfigDir, 0600 where that means something); the system's
+  CA store, `--ca` for another.
+- **Gateway endpoints** (client listener, person's key):
+  `GET /v1/kb/collections` (what this key may read/write),
+  `GET /v1/kb/collections/{id}/manifest`, `PUT
+  /v1/kb/collections/{id}/documents` (one document per request: name,
+  digest, bytes, chunks with text and embedding; dims checked, chunk count
+  and size bounded), `DELETE …/documents?name=`. Write needs a write grant
+  or ownership on the key's chain (`CollectionsFor(…, write)`); otherwise
+  404.
+- **Administration** stays with operators: `llmgw ctl kb
+  create|list|delete|grant|revoke|docs`, admin API `/api/v1/kb`, audit,
+  console screen; `docs/knowledge.md`.
+- Trade-off accepted: vectors come from the client, so a writer could store
+  vectors that do not match their text — only in collections they may
+  already write.
+Built: `internal/kb/extract` (formats above; helpers found on PATH and in
+each system's install places), `internal/kb/chunk` (≈1800 characters under
+the heading path, a paragraph of overlap, sections kept apart once a chunk
+is a third full), `internal/kb/kbsync` (ledger in SQLite at
+os.UserConfigDir/llmgw-kb/ledger.db: documents keyed by gateway, collection
+and name with file, size, mtime, digest, state, reason; runs), `cmd/llmgw-kb`
+(login, collections, sync, status, history, formats), `internal/router/kb.go`
+(the four gateway endpoints, `X-Llmgw-Project` for personal keys, audit),
+`internal/control/adminapi/kb.go` + `llmgw ctl kb` + console Knowledge
+screen, `docs/knowledge.md`; CI builds the client for all six targets and
+releases attach it. Tested: every format from generated fixtures and a
+Turkish PDF (OCR path with Tesseract here), sync against a fake gateway and
+end to end against the router on pgvector.
+Tried on the PoC 2026-10-08 (control planes 140 and 150 on a dev build of
+this branch; their previous binary is /usr/local/bin/llmgw.pre-rag): Ubuntu's
+pgvector 0.6.0 installed on 149/150 and enabled; bge-m3's weights were never
+in the mirror (fetched from CompendiumLabs/bge-m3-gguf@1864f47f) and it is
+held at 1 replica; tree Demo/ik/belgeler and Demo/satis/crm with a key each;
+collection Demo/politikalar filled from a Mac through an SSH tunnel with ten
+Turkish files in nine formats (9 sent in 5 s, the .doc skipped for want of
+LibreOffice; a rerun sent nothing in 0 s; a changed file alone was sent).
+Nine Turkish questions each found their page first once search scored by
+similarity plus word coverage (rank fusion had got two wrong). The other
+team's key saw no collection and got 404 by id; with a read grant it read
+and could not write; revoked, nothing again. On the live store the llmgw
+role (not superuser, no BYPASSRLS) sees 0 chunks with no scope set and 0
+with the organization alone.
+Left: the gateway's collection list shows the owner as team:ID, not its
+path; no LibreOffice on the dev machine, so the legacy
+Office path is untested beyond its error; the client is unsigned (Gatekeeper
+and SmartScreen warn); Outlook `.msg` is not read; one document over ~60 MiB
+of JSON (a few thousand pages) is refused rather than split.
+PostgreSQL tests of store, router and adminapi share LLMGW_TEST_POSTGRES and
+the store's TRUNCATE each test; run them with `go test -p 1` or one package
+at a time, or a router test loses its rows mid-run.
+
+### F3. Retrieval on the request path (M) — done 2026-10-08 (branch `rag`)
+- `llmgw.knowledge: [collections]` in the body (stripped before the
+  engine), or default collections per project/key.
+- Allowed set from the caller's chain (§7.9.2), cached like limits; grants
+  add up down the tree. Personal keys and sessions need `x-llmgw-project`
+  (or the playground's picker) naming a scope the person is a member of.
+- A collection not allowed answers 404, exactly as a missing one.
+- Hybrid search with the filter inside the query (per collection, its own
+  partial index; pgvector 0.6 is enough —
+  scored by similarity plus the share of the question's words held (built
+  in F1/F2 as SearchChunks), `/v1/rerank` when a reranker is in the catalog. Top
+  chunks into a system message with document and position;
+  `x-llmgw-sources` lists them.
+- Leaks closed: answer-cache key includes the allowed set; per-organization
+  `cache_salt` on RAG requests; injected passages under `no_store` and the
+  guard's prompt-injection check; returned doc ids in the audit.
+- Done: router tests that tenant B never receives tenant A's chunks
+  (including via cache), that a personal key without a project is refused,
+  metrics (retrieval latency, chunks returned), docs.
+
+Built (`internal/router/rag.go`): `X-Llmgw-Knowledge` (header, as the
+other llmgw extensions are, not a body field), `X-Llmgw-Knowledge-K`;
+collections from the key's chain (kbEffective, shared with the kb
+endpoints), 404 for one not readable; the question embedded through the
+ordinary path on the caller's principal; SearchChunks; passages trimmed to
+the model's context; added to the one leading system message (or a new
+one); `x-llmgw-sources`; before guardInput; never cached; metrics. Default collections built (kb_defaults on org/team/project/key, narrowest
+wins, 30 s cache, `ctl kb default`, console action; never refuse); the
+portal chat's Documents picker (collection, default or none; sources listed
+under the answer; a header the page sends itself is dropped). `cache_salt` per organization on requests given passages, sent to vLLM
+replicas only (chosen replica's engine). Checked on the PoC 2026-10-08:
+vLLM 0.30.0 (cuda01) with TinyLlama-1.1B AWQ run by hand beside the 35B in
+the card's last 2.3 GB — a 668-token prompt reused 656 tokens under the same
+salt and 0 under another's, with identical text (llmgw cannot place it
+there: placement takes whole GPUs). Live on the PoC: a team default answers
+a request without the header, `none` turns it off, an explicit header with
+k=2 gives two sources and no more, llmgw-kb lists owners by path. Left: returned passages
+are not audited per request (the header and the kept request record them).
+
+### F4. Adapter training (L, after F1–F3) — done 2026-10-08 (branch `adapters-f4`)
+A run-to-completion job placed by the control plane like a replica and
+supervised by the agent like an engine, in a Python environment the agent
+installs (TRL + PEFT, QLoRA). Data: Q/A pairs generated from a collection by
+a cluster model through `/v1/batches`, reviewed by people. Output to the
+mirror, into the model's `adapters`, through canary and a quality suite.
+Runs when the scaling schedule frees a GPU; 8–14B dense bases on one card.
+Never train on a collection some users of the adapter may not read.
+Tried 2026-10-08 on the PoC, by hand: Qwen2.5-0.5B-Instruct (fp16) with a
+LoRA (r=16, every linear layer, 8.8M trainable), trained beside the 35B in
+cuda01's last 2.3 GB (peak 2158 MiB on the card, 64 s for 138 steps, PEFT in a
+venv over the vLLM engine's torch). Data: 90 Turkish Q/A pairs a cluster
+model wrote from the nine Demo/politikalar documents, through the gateway.
+On nine questions it had not seen, phrased otherwise: base 0/9, 6 epochs
+4/9, 4 epochs at a lower rate 2/9. Wrong answers mixed facts across
+documents (the meal card's amount, the office day). The same nine through
+retrieval: 9/9 with sources. So F4 is for format, vocabulary and tone, and
+facts come from F1–F3; a bigger base on a free card would do better but not
+change that. The training itself fits a small slot on a shared card, which
+the job design can rely on for small bases.
+
+Built: `ctl kb dataset` (POST /api/v1/kb/collections/{id}/dataset, NDJSON
+with each pair's collection); train_jobs/train_artifacts (migrations 0037,
+0038 owner); internal/control/train (queue, one job per node, artifacts in
+1 MiB pieces, a node that goes fails its job); internal/agent/train (the
+newest vLLM's python3 + `uv pip install --no-deps --target` peft 0.21.2 and
+accelerate 1.15.0; llama.cpp's conversion code for the installed release,
+untarred from its source; the card with the most free memory, if enough;
+runs as llmgw-engine); `ctl train start|list|show --follow|cancel|rm|
+download|publish`; the mirror's PUT /v1/models/{model}/{version}/{file};
+publish uploads the GGUF (llama.cpp) or PEFT files (vLLM) and sets the
+adapter after checkAdapters. Owners: catalog.Adapter.Owner, enforced in the
+router (404, hidden from /v1/models); a job's owner defaults to its single
+collection's owner, must be able to read every collection the pairs came
+from, and pairs from a knowledge base are never public. Tried on the PoC:
+dataset of 90 pairs in 89 s; a job trained beside the 35B (peak 2214 MiB,
+138 steps in about a minute) and sent back PEFT and GGUF.
+Left: a job dies with its agent's stream (no resume); no console screen for
+training yet; quality gating of a published adapter is by hand (ctl eval).
+
+---
+
+### F5. Large imports (M) — measured 2026-10-09, open
+The metallurgy corpus (2,314 files, 108,779 chunks, bge-m3 on the one
+GPU) went in at about 25 chunks/s, the same over a phone hotspot and the
+LAN. The engine's slots were the limit: llama.cpp `--parallel 4` embeds
+four texts per pass, while llmgw-kb sends 32 a request from 6 workers. The
+usage rows show 4.6 s a request on average, 1.8 s of it in the gateway's
+queue and 0.04 s in the guard. The GPU sat at 18–50% and the store machine
+at a third of its CPUs.
+- An embedding model's profile with many slots by default (`--parallel`
+  32–64, `--ctx-size` to match, `--batch-size`/`--ubatch-size` at least the
+  longest chunk), or `ctl model serve` choosing that for a model with
+  `capabilities.embeddings`. With the default 512, one 516-token chunk failed.
+- Embedding on the gateway's side of a document PUT, so the client sends
+  text only: today an embeddings answer is ~21 KB of JSON a chunk, and
+  the text and vector go up again (~11 KB).
+- The store next: HNSW inserts took half of one of pg01's two cores; at a
+  million chunks the vectors alone outgrow its 4 GB. halfvec (pgvector ≥
+  0.7) and a larger store machine.
+- Done: a batch of embeddings checked text by text against the context
+  (branch `embeddings-batch`); it was refused as one over-long prompt.
+- Done: `query_prefix` on an embedding model, before questions only.
+- Qwen3-Embedding-4B on the RTX 6000 (2026-10-09): llama.cpp Q8 1,575
+  tok/s as first set up (32k cache, 16 slots), 2,700 with a 16k cache,
+  FlashAttention and ubatch 1–2k; f16 no better. vLLM fp16 (`--runner
+  pooling`) 6,200 direct, 4,700–4,900 through the gateway in the import,
+  and its vectors match llama.cpp's (cosine 0.9995). The 103,579 chunks
+  took 3 h. Wanted: an embedding model's GPU profile on vLLM and CPU
+  profile on llama.cpp, which needs a source per profile (the catalog has
+  one per model; the import swapped the entry by hand), and an admission
+  queue limit that suits bulk imports (30 s turned some away).
+- Retrieval on 40 questions written from the corpus, document in the top
+  5: bge-m3 31/40 (Turkish questions 13/21), Qwen3-Embedding-4B 37/40
+  (18/21), with its query prefix 36/40 and first place 31/40 instead of
+  28 (MRR 0.738 → 0.789 → 0.838). Next: translating the question with the
+  chat model, and a reranker.
+
 ## D. Facts worth knowing before starting
 
 - Admin API: `127.0.0.1:9090`, no auth, `ctlClient` in `internal/cli/ctl.go`.

CHANGELOG.md

@@ -17,11 +17,325 @@ is only released once its **exit criterion** is met, not when its code lands.
 | v0.10.0 | — | InfiniBand fabrics and NVLink, doctor, node add, engines installed by the nodes | released |
 | v0.11.0 | — | High availability from setup, the setup and console screens, usage sums and per-tenant alerts | released |
 | v0.12.0 | — | The cluster's settings in the store, in force when changed; the firewall switch; wipe | released |
+| v0.13.0 | — | The whole system stopped and started from one command; Qwen3Guard judges, on a GPU and CPUs together | released |
+| v0.14.0 | — | Knowledge bases: documents from people's own machines, kept apart per tenant, answered from | released (v0.14.1: adapters, the tree over the API, the firewall on every machine) |
 | v1.0.0 | 5 | Tenancy, prefix affinity, benchmarking | planned |
 
 Until v1.0.0 the minor version is the breaking one: configuration formats and Go
 APIs change between minors without a deprecation period.
 
+## Unreleased
+
+## v0.14.1 — 2026-10-09
+
+Adapters trained on the nodes from a knowledge base and kept private to
+their owners; the cluster's and the agents' settings edited in one place;
+the tenancy tree and keys over the admin API and in Terraform; knowledge
+bases that hold up with many documents; the firewall setting on every
+machine.
+
+**Upgrading.**
+- **Nodes** need this release to follow `agents.*` in the cluster's
+  settings; until a node is upgraded its agent.yaml is what it runs on.
+- **A separate mirror machine and store members that run nothing else**
+  follow the firewall setting once what reads it is installed: run `sudo
+  llmgw mirror init` on the mirror again with the flags it was set up with,
+  and `llmgw server init --patroni` again for the store's members. A control
+  plane set up before v0.12.0 needs `sudo llmgw firewall sync --install`.
+- **The store's PostgreSQL memory** applies after each member is restarted,
+  one at a time, the leader last; until then it runs as before.
+- The store gains tables and columns; nothing to do. Existing knowledge-base
+  collections get their full-text index when a control plane starts, and
+  keep `vector` storage; only new collections use `halfvec`, and only on
+  pgvector 0.7 or later.
+- Cached knowledge-base passages in vLLM are made again once, under the new
+  `cache_salt`.
+
+### Added
+
+- **The cluster's settings on the portal** (BACKLOG E4): `/admin/settings`
+  for administrators. It edits the document `llmgw ctl config edit` opens:
+  checked as a control plane would load it, refused when made from a
+  version since replaced, with the history to view and roll back. Changes
+  are audited by size, never by contents.
+- **The nodes' agent settings in the cluster's settings** (BACKLOG E3):
+  `agents.default` and `agents.nodes.<id>` hold everything agent.yaml can
+  say but the agent's id, control planes, address and certificate. They are
+  sent to the agents as they change; an agent starts again on new settings,
+  or on an edited agent.yaml, with its engines serving throughout. Settings
+  that would not load are refused on save, and set aside on a node rather
+  than keep its agent from starting. Nodes need this release to follow
+  them.
+- **LoRA adapters trained on the nodes** (BACKLOG F4): `llmgw ctl kb dataset`
+  writes question/answer pairs from a collection with a model of the
+  cluster. `llmgw ctl train start` trains an adapter on a node with its
+  installed vLLM's torch and a small overlay, on a card with enough memory
+  free (beside an engine, if it leaves that much), converting to GGUF for
+  llama.cpp. `llmgw ctl train publish` keeps it on the mirror (which now
+  accepts uploads) and makes it the model's adapter.
+- **Adapters with owners.** An adapter's `owner` (org, team, project, user
+  or key) is the only part of the tree that sees it; to anyone else it is a
+  model that does not exist. An adapter trained on a knowledge base belongs
+  to a part of the tree that may read it, and is never made public.
+- The knowledge bases' vLLM `cache_salt` follows the collections searched,
+  not the organization, so two teams of one organization no longer share
+  cached passages.
+- **Organizations, teams, projects, API keys and memberships in the admin
+  API**: `/api/v1/orgs` (the tree), `POST /api/v1/teams` and `projects`,
+  `PATCH` to rename or move, `DELETE` for what is empty; `/api/v1/keys` to
+  issue (the key is in the answer, once), list and revoke; `/api/v1/users`
+  and `/api/v1/users/{id}/memberships`. They were only `llmgw org`, `keys`
+  and `users` on a control plane before. The Terraform provider manages them
+  as `llmgw_org`, `llmgw_team`, `llmgw_project`, `llmgw_key` (the key in a
+  sensitive `secret`) and `llmgw_membership`.
+- **The firewall setting reaches every machine.** A separate mirror and the
+  store's members follow `firewall.enabled` and `firewall_networks` as nodes
+  and control planes did: the leading control plane tells the mirror over
+  its API, and keeps the state in the store for members that run no llmgw,
+  which a timer on each reads. The mirror also lets in the network of every
+  node the leader knows. `mirror init` and `server init --patroni` install
+  what follows it; run them again on a cluster set up before.
+
+### Changed
+
+- **The store's PostgreSQL is sized to its machine's memory.** Before, it ran
+  on PostgreSQL's defaults (128MB cache) under Patroni, or on 96MB on one
+  machine. Now each member gets a share of its own RAM, a smaller one where
+  a control plane or agent shares the machine, and the share is read again
+  every time `server init --patroni` runs. The new cache is applied by
+  restarting one member at a time, the leader last.
+- **Knowledge-base searches hold up with many documents.** New collections
+  store their index as `halfvec` where pgvector is 0.7 or later (half the
+  memory, and embeddings up to 4000 dimensions instead of 2000); older
+  collections, and clusters on pgvector 0.6, keep `vector`. Each collection
+  has its own full-text index, as it has its own vector index, made for
+  existing collections when the control plane starts. The text arm ranks by
+  the cheap measure first and scores only those candidates.
+
+### Fixed
+
+- A knowledge-base search asking for more than 10 passages
+  (`X-Llmgw-Knowledge-K`) got at most 40 candidates per collection from the
+  vector index, whatever it asked for: pgvector's `hnsw.ef_search` is now
+  raised to the number wanted, up to 1000.
+- `llmgw server backup` and `restore` against a replicated store over TLS
+  failed looking for `~/.postgresql/root.crt`: pg_dump was given
+  `sslmode=verify-full` without the cluster's authority.
+
+## v0.14.0 — 2026-10-08
+
+Knowledge bases: a company's documents, put in by people from their own
+machines, kept apart per tenant, and answered from.
+
+**Upgrading.**
+- **pgvector on a cluster set up before this release.** Knowledge bases
+  need pgvector, which `server init` and the high-availability database
+  setup now install, but a rerun on a running cluster does not:
+  1. Install `postgresql-<version>-pgvector` on every database machine.
+  2. As postgres, run `CREATE EXTENSION vector` in llmgw's database.
+  3. Restart the control planes.
+  Until then everything else works, and `llmgw ctl kb list` says why
+  knowledge bases are off. On a SQLite store they stay off.
+- **An embedding model.** A collection needs a catalog model with
+  `capabilities.embeddings`, for example bge-m3, whose weights are in the
+  mirror.
+- **The client.** People download `llmgw-kb` for their system from this
+  release.
+
+### Added
+
+- **Collections and who may read them.** A collection is owned by an
+  organization, team, project or person. Grants add up down the tenancy
+  tree and never cross an organization. A collection a key may not read is
+  answered as one that does not exist. Chunks are kept in PostgreSQL with
+  pgvector, partitioned by organization, with an HNSW index per collection,
+  and row-level security on the table as a second line behind every
+  query's own filter. A project, team or organization that owns collections
+  is not deleted; a person's collections go with their account.
+- **`llmgw-kb`**, its own download for Windows, macOS and Linux, run by
+  people with their own key. It reads documents where they are and sends
+  text and vectors, never the files, for what changed only; a SQLite ledger
+  on the machine remembers what it did.
+  - It reads PDF, Word, Excel, PowerPoint, OpenDocument, RTF, EPUB, HTML,
+    MHTML, mail with attachments, notebooks, subtitles, ZIP archives, and
+    text and source code in any common encoding.
+  - Legacy Office files are read with LibreOffice, and scanned pages with
+    Tesseract, when those are installed.
+  - `--dry-run` says what would leave the machine; `--delete` removes only
+    what that machine sent.
+- **Answers from documents.** A chat request names collections in
+  `X-Llmgw-Knowledge`, or its part of the tree has a default (`llmgw ctl kb
+  default`). The best passages, chosen by meaning plus a little for the
+  question's words, are added to the system message, numbered, and
+  `x-llmgw-sources` says which were used. This happens before the
+  guardrails judge the request, and the answer is never served from the
+  answer cache. On vLLM replicas the organization's `cache_salt` keeps
+  tenants' prefix caches apart. It works through the Responses and Messages
+  APIs too.
+- **Administration**: `llmgw ctl kb` (create, list, grant, revoke, default,
+  docs, rm, search, delete), `/api/v1/kb`, the console's Knowledge screen,
+  and a Documents picker on the portal's chat page.
+- Metrics `llmgw_kb_retrievals_total` and `llmgw_kb_retrieval_seconds`;
+  [docs/knowledge.md](docs/knowledge.md).
+
+## v0.13.0 — 2026-10-08
+
+The whole system stopped, started and restarted from one command, a
+guardrails judge that no longer refuses ordinary coding requests, and the
+rest of the field report from v0.10.6.
+
+**Upgrading.** A cluster keeps the judge its policies name: one installed
+before this release goes on asking Llama Guard 3 1B until `llmgw ctl guard
+judge use qwen3guard-0.6b`, which the leader then fetches and serves by
+itself, and which replaces the policies' Llama Guard category codes.
+
+### Added
+
+- **A judge on a GPU and one on CPUs together**: a policy's `judge.models`
+  is a list asked in turn, the first that answers giving the verdict; the
+  next is asked when one has no replica serving, fails, or misses its
+  `latency_budget`. `llmgw ctl guard judge use qwen3guard-4b
+  qwen3guard-0.6b` sets the order for every policy with a judge, and
+  `guard judge status` says which judge stands in for which. A judge on a
+  GPU is served as a `standard` model, so it takes a card only from what
+  may give one up.
+- **A judge a policy names is fetched if it is missing**: the leader
+  fetches the weights of a judge llmgw knows (`qwen3guard-0.6b`, `-4b`,
+  `-8b`, `llama-guard-3-1b`, `-8b`) into the mirror and serves it, when a
+  policy names it and the catalog does not have it — and again if it is
+  removed. `guard judge enable --gpu` serves Qwen3Guard-Gen 4B on the class
+  with the most free cards.
+- **The whole system stops, starts and restarts from one command**: `llmgw
+  stop` takes every model down and answers every request for one with 503
+  (`system_stopped`), keeping the control planes, agents and mirror running
+  and no model's definition changed; `llmgw start` brings every model back.
+  It is the cluster's setting `serving.stopped`, live on every control plane.
+  `llmgw restart` restarts every node's agent and every control plane, the
+  leader last, with the engines serving throughout. `llmgw status` says
+  whether the system serves. One machine's services are `llmgw service
+  status|start|stop|restart`.
+- **A replica says what it computes on**: `ctl replicas` (ON), the console
+  and the admin API show `CUDA` with the GPUs its processes hold memory on
+  and how much, as nvidia-smi sees them, or `CPU` — for every engine, kept by
+  the agent across its own and the control plane's restarts. llama.cpp's
+  layer count is added when it says it, which it does only at `-lv 4`.
+- **`ctl engine-check llamacpp` computes a token on the GPU**: a small GGUF
+  (`--model`, or the smallest in the weights cache) is loaded with every layer
+  on the GPU at `-lv 4` and asked for one token; a build whose CUDA backend
+  does not load fails, by llama.cpp's own account of its layers. With no
+  `--bin` it checks the build the agent installed.
+- **Engine versions and compute capability on nodes**: `ctl nodes` shows
+  `llamacpp b10991` and `6000-24g (7.5)`; the admin API gives
+  `engine_versions` and `compute_capabilities`. A profile's `engine_version`
+  restricts it to nodes whose engine version starts with it, and
+  explain-placement says which version a node has, or which engines.
+- **`ctl restart --max-surge N` and `--max-unavailable N`**: that many of a
+  model's replicas restart at once, beside themselves or stopped first.
+- **Multi-node readiness apart from the node's**: doctor gives both, with
+  fabric findings in a section of their own; `node add` says the multi-node
+  one; `ctl nodes` shows a declared fabric's adapters that are down.
+- **A forgotten password is set through a link**: `/forgot` mails one when
+  the cluster's `mail` settings say how (SMTP, sender, the public
+  `base_url` it points at), saying the same whether or not the address has an
+  account; `/reset` sets the password once and closes every session. Without
+  mail, `llmgw users reset-link` makes the link for an operator to hand over.
+
+### Fixed
+
+- **A replica restarted in place keeps its room**: the replica was stopped
+  and the pass went on to other models before placing its replacement, so a
+  model waiting for room took the freed cards. On the reference cluster a
+  GPU judge waiting for the only card took it from the model being
+  restarted there. The replacement is now placed in the same step.
+- **The judge no longer refuses ordinary coding requests**: Llama Guard 3
+  1B, the judge `server init` installed, could not tell "continue", "devam
+  et" or "how do I kill all child processes in bash" (0.51 to 0.92 unsafe)
+  from harmful requests (0.97 to 1.00), at any quantisation. The judge is
+  now Qwen3Guard-Gen 0.6B (format `qwen3guard`), which gave the same
+  ordinary requests 0.15 at the most and the harmful ones 1.00, on CPUs,
+  in 0.8 GB. Its third verdict, controversial, counts as safe unless the
+  policy says `controversial: unsafe`. `guard judge use qwen3guard-0.6b`
+  moves an installed cluster to it, and replaces Llama Guard's category
+  codes, which Qwen3Guard never names, with the baseline's.
+- **A long message is judged rather than refused**: a judge read the whole
+  last user message, and one with a file pasted into it overflowed the
+  judge's context, so it gave no verdict and `fail: closed` refused the
+  request. It now reads the start and the end, about 6000 characters.
+- **A replica no longer stays "starting" while it serves after a restart**:
+  an agent that registered before its adopted engines had answered their
+  health check never said they had, so the control plane kept them starting
+  and routed nothing to them — on the reference cluster, after `llmgw
+  restart`, the only replica of a model. The agent now reports each shard
+  the registration called not ready once it is.
+- **A failover no longer replaces healthy replicas**: a standby that took
+  over placed at once, before the agents had reconnected and said what runs,
+  so it started a second copy of replicas it did not know of and then drained
+  the healthy ones as surplus — on the reference cluster the guardrails'
+  judge and a model's replica, within twenty seconds of a leader restarting.
+  Placement now waits after a takeover as it does at start.
+
+### Changed
+
+- **A GPU placement llama.cpp says it cannot compute on is refused** when it
+  warns that it found no usable GPU, even where nvidia-smi cannot be asked;
+  where it can, holding no memory on its GPUs refuses it as before.
+
+## v0.12.2 — 2026-10-06
+
+The guardrails turned off and on as a whole, and a judge that refuses
+ordinary requests explained: the hazard it saw is told to the client and
+kept with the request, and `llmgw ctl guard judge ask` shows what it says of
+any text. Nothing to do to upgrade; the store gains a column.
+
+### Added
+
+- **The guardrails turn off and on as a whole**: `llmgw ctl guard off` /
+  `on`, which is `guard.enabled` in the cluster's settings — in force within
+  seconds, the policies and their attachments kept. `ctl status`, `ctl guard
+  list` and the guardrails page say when they are off.
+- **`llmgw ctl guard judge ask TEXT`** has the judge read a text and shows
+  its answer as given and as read — for finding out why ordinary requests
+  are refused. Nothing of the text is kept, not even in the audit.
+
+### Changed
+
+- **A judge's refusal names the hazard**: the client is told `(judge:
+  S6,Specialized Advice)` rather than `(judge)`, and the verdict is kept with
+  the request (`guard_rule`): `ctl request`, the server's log and the
+  guardrails page show it. Llama Guard 3's codes come with their names,
+  S9 Indiscriminate Weapons, without a categories list.
+
+## v0.12.1 — 2026-10-06
+
+The guardrails' judge, served and watched without server init: a judge
+missing, served as the wrong kind of model, or not answering is said on
+`ctl status` and the console, and `llmgw ctl guard judge enable` puts it
+right. Nothing to do to upgrade.
+
+### Added
+
+- **`llmgw ctl guard judge status` and `enable`**: each judge model the
+  policies ask, whether it is in the catalog, as a judge, with replicas
+  ready and answering — and serving the missing one, Llama Guard 3 1B on
+  CPUs under the name the policies give it, with `--replace` for a model of
+  that name that is not a judge. No more `server init` to put a judge back.
+  Also **Do → Serve the guardrails' judge** in the console.
+- **A missing or broken judge is said where it is looked for**: `llmgw ctl
+  status` and the console's Overview list each judge with a problem — not in
+  the catalog, not a judge, no replica ready, or failing to answer — and
+  what that does to requests (refused or unjudged).
+
+### Fixed
+
+- **A Llama Guard file served from the mirror page is the judge**: it was
+  added as a model clients use, under whatever name the form had, so the
+  policies never found it and refused everything they judged
+  (`guard_unavailable`). It is now added as a judge, under the name the
+  policies ask.
+- **`ctl` says so when the API answers with a page**: a command a control
+  plane does not know — the leader on an older release, through a standby —
+  failed with "invalid character '<'"; it now says which and why.
+
 ## v0.12.0 — 2026-10-06
 
 A minor version, because where the configuration lives changes: everything in

Makefile

@@ -6,6 +6,7 @@ LDFLAGS := -X '$(MODULE)/internal/version.Version=$(VERSION)'
 
 build:
 	go build -ldflags "$(LDFLAGS)" -o bin/llmgw ./cmd/llmgw
+	go build -ldflags "$(LDFLAGS)" -o bin/llmgw-kb ./cmd/llmgw-kb
 
 test:
 	go test ./...

README.md

@@ -3,8 +3,9 @@
 A single-binary LLM gateway for clusters with multiple InfiniBand fabrics,
 NVLink domains and heterogeneous GPUs.
 
-It serves LLMs running on your own GPU servers (vLLM / SGLang) behind one
-OpenAI-compatible endpoint. Design document: [llmgw-plan.md](llmgw-plan.md).
+It serves LLMs running on your own GPU and CPU servers (vLLM, SGLang,
+llama.cpp, TensorRT-LLM) behind one OpenAI- and Anthropic-compatible
+endpoint, and answers from your own documents. Design document: [llmgw-plan.md](llmgw-plan.md).
 
 ## Getting started
 
@@ -37,6 +38,9 @@ Releases track the roadmap phases; see [CHANGELOG.md](CHANGELOG.md).
 | v0.9.0 | — | Managing people and the tenancy tree | released |
 | v0.10.0 | — | InfiniBand fabrics and NVLink, doctor, node add, engines installed by the nodes | released |
 | v0.11.0 | — | High availability from setup, the setup and console screens, usage sums and per-tenant alerts | released |
+| v0.12.0 | — | The cluster's settings in the store, in force when changed; the firewall switch; wipe | released |
+| v0.13.0 | — | The whole system stopped and started from one command; Qwen3Guard judges, on a GPU and CPUs together | released |
+| v0.14.0 | — | Knowledge bases: documents from people's own machines, kept apart per tenant, answered from | released (v0.14.1: adapters, the tree over the API, the firewall on every machine) |
 | v1.0.0 | 5 | Tenancy, prefix affinity, benchmarking | planned |
 
 As of v0.6.0 a cluster runs unattended: agents report their hardware, the
@@ -44,38 +48,67 @@ control plane places replicas — including ones larger than a single machine
 models load when someone asks for them and stop when nobody does, and keys,
 budgets and rate limits decide who may use any of it.
 
-Since v0.11.0, not yet released (see [CHANGELOG.md](CHANGELOG.md),
-"Unreleased"):
+Since v0.11.0 (each release's detail is in [CHANGELOG.md](CHANGELOG.md)):
 
-- **Security hardening** — signed releases only, a token on the admin API,
-  engines not run as root, TLS for the agent port, etcd and Patroni, a
-  firewall for llmgw's internal ports, encrypted backups, bounded
-  connections, sign-in throttling, a Content-Security-Policy.
-- **Guardrails** — policies that check requests and answers (PII, secrets,
-  prompt injection, deny patterns, a judge model such as Llama Guard), in
-  observe or enforce mode, attached to any part of the tenancy tree:
+- **Adapters, and the tenancy tree over the API** (v0.14.1). `llmgw ctl
+  kb dataset` and `llmgw ctl train` make a LoRA adapter from a collection
+  on a node's spare GPU memory, kept on the mirror and seen only by its
+  owner. Organizations, teams, projects, keys and memberships are in the
+  admin API and the Terraform provider. The agents' settings are edited
+  with the cluster's, on the portal too, and the firewall setting reaches
+  every machine.
+- **Knowledge bases** (v0.14.0). People put a company's documents in from
+  their own machines with `llmgw-kb`, a separate download for Windows,
+  macOS and Linux that reads PDF, Office, OpenDocument, mail and many more
+  formats and sends text and vectors, never the files. Collections are kept
+  apart per tenant, down to row-level security in PostgreSQL. A chat
+  request, or a part of the tree by default, is answered from them with
+  its sources: [docs/knowledge.md](docs/knowledge.md).
+- **One command for the whole system** (v0.13.0): `llmgw stop`, `start` and
+  `restart`. Qwen3Guard is the guardrails' judge, on a GPU and on CPUs
+  together.
+- **Settings that apply when changed** (v0.12.0). The cluster's
+  configuration lives in the store, the same on every control plane and in
+  force within seconds (`llmgw ctl config`, the console's Settings). The
+  firewall can be switched on and off for the whole cluster.
+- **Security hardening** (v0.11.2): signed releases only, a token on the
+  admin API, engines not run as root, TLS for the agent port, etcd and
+  Patroni, a firewall for llmgw's internal ports, encrypted backups,
+  bounded connections, sign-in throttling, a Content-Security-Policy.
+- **Guardrails** (v0.11.2). Policies check requests and answers for PII,
+  secrets, prompt injection, deny patterns, and with a judge model; they
+  observe or enforce, and attach to any part of the tenancy tree:
   [docs/guardrails.md](docs/guardrails.md).
-- **Administration** — an audit log of every change (`ctl audit`), staff
-  roles (viewer, operator, billing, admin), single sign-on through OpenID
-  Connect and provisioning through SCIM 2.0, key lifetime limits and
-  rotation with overlap (`keys ttl`, `keys rotate`), usage export as CSV
-  (`ctl usage export`).
-- **Notifications** — webhooks, signed or Slack-shaped, for a model parked,
-  a node lost, a budget spent and new weights published (`notify.webhooks`,
-  `ctl notify test`), and Alertmanager receiver templates.
-- **Traffic and models** — request tags, canaries of new weights with
-  automatic rollback, quality suites gating updates, a response cache,
-  kept and replayed requests, batches, Anthropic's Messages API, more of
-  OpenAI's (structured output, images, rerank, audio, moderations),
-  scaling by the clock, SLO-aware admission, TensorRT-LLM, MIG, models
-  imported from Ollama, routing by what each vLLM replica's cache holds,
-  LoRA adapters loaded on demand, other llmgw clusters as providers, and
-  placement that knows racks, PDUs and power caps.
-- **Operations** — `server restore` and off-machine backup copies, secrets
-  in Vault, forgetting a person and keeping a project's requests in the
-  cluster, offline installs from a bundle, a Helm chart, tracing from the
-  control plane to the engines, the admin API in OpenAPI with a Terraform
-  provider, and the portal in English and Turkish with its docs at /docs.
+- **Administration** (v0.11.2):
+  - an audit log of every change (`ctl audit`);
+  - staff roles (viewer, operator, billing, admin);
+  - single sign-on through OpenID Connect and provisioning through SCIM
+    2.0;
+  - key lifetime limits, and rotation with overlap (`keys ttl`, `keys
+    rotate`);
+  - usage export as CSV (`ctl usage export`).
+- **Notifications** (v0.11.2): webhooks, signed or Slack-shaped, for a
+  model parked, a node lost, a budget spent and new weights published, and
+  Alertmanager receiver templates.
+- **Traffic and models** (v0.11.2):
+  - request tags, canaries of new weights with automatic rollback, and
+    quality suites gating updates;
+  - a response cache, kept and replayed requests, and batches;
+  - Anthropic's Messages API, and more of OpenAI's (structured output,
+    images, rerank, audio, moderations);
+  - scaling by the clock, and SLO-aware admission;
+  - TensorRT-LLM, MIG, and models imported from Ollama;
+  - routing by what each vLLM replica's cache holds, and LoRA adapters
+    loaded on demand;
+  - other llmgw clusters as providers, and placement that knows racks,
+    PDUs and power caps.
+- **Operations** (v0.11.2):
+  - `server restore`, off-machine backup copies, and secrets in Vault;
+  - forgetting a person, and keeping a project's requests in the cluster;
+  - offline installs from a bundle, and a Helm chart;
+  - tracing from the control plane to the engines;
+  - the admin API in OpenAPI, with a Terraform provider;
+  - the portal in English and Turkish, with its docs at /docs.
 
 Every one of these is in the console as well as on the command line:
 [docs/console.md](docs/console.md).
@@ -124,6 +157,9 @@ Every page is also in the binary: `llmgw docs`, or **Manual** in the console.
   official OpenAI and Anthropic SDKs, and llmgw's extensions
 - [docs/guardrails.md](docs/guardrails.md) — checking requests and answers:
   policies, detectors, a judge model, and where each applies
+- [docs/knowledge.md](docs/knowledge.md) — knowledge bases: collections,
+  who may read them, `llmgw-kb` on people's machines, and answers from
+  documents
 - [docs/choosing-models.md](docs/choosing-models.md) — which models to run,
   how many of each, and where the weights should live
 
@@ -152,7 +188,7 @@ Every page is also in the binary: `llmgw docs`, or **Manual** in the console.
 ## Building
 
 ```sh
-make build      # bin/llmgw
+make build      # bin/llmgw and bin/llmgw-kb
 make test       # unit tests
 make race       # with the race detector
 ```
@@ -169,6 +205,7 @@ llmgw ctl cordon gpu-a-01            # stop new placement, leave what runs
 sudo llmgw doctor                    # on a node: what its GPUs, NVLink and fabric need
 llmgw ctl audit --since 24h          # who changed what
 llmgw ctl usage export --from 2026-09-01 --to 2026-10-01 --by project,model
+llmgw ctl kb list                    # knowledge-base collections, grants, defaults
 ```
 
 Bare `llmgw` opens the console, which does all of this from one screen.

api/proto/llmgw/v1/agent.proto

@@ -28,6 +28,8 @@ message AgentMessage {
     FabricTestReply fabric_test = 6;
     UpgradeReply upgrade = 7;
     KVEvents kv_events = 8;
+    TrainStatus train_status = 9;
+    TrainArtifact train_artifact = 10;
   }
 }
 
@@ -72,6 +74,10 @@ message ControlMessage {
     UpgradeAgent upgrade = 10;
     ControlPlanes control_planes = 11;
     Firewall firewall = 13;
+    RestartAgent restart_agent = 14;
+    AgentSettings agent_settings = 15;
+    TrainStart train_start = 16;
+    TrainCancel train_cancel = 17;
   }
   // trace is the W3C trace context (traceparent, and tracestate when there
   // is one) of the control plane's span that sent this command, so what the
@@ -81,6 +87,12 @@ message ControlMessage {
   map<string, string> trace = 12;
 }
 
+// RestartAgent has the agent start again, as llmgw restart asks of every
+// node: its engines keep serving and are taken over again.
+message RestartAgent {
+  string reason = 1;
+}
+
 // Firewall is whether llmgw's firewall is kept on the node — its nftables
 // table, which lets the cluster's networks alone reach the engines' ports and
 // the node exporter — sent when an agent registers and whenever the
@@ -96,6 +108,17 @@ message Firewall {
   int32 engine_ports_to = 4;
 }
 
+// AgentSettings is what the cluster's settings say of this node's agent
+// (agents.default, with agents.nodes.<id> over it), sent when it registers
+// and whenever they change. The agent keeps it beside agent.yaml, which
+// holds only what it needs to start; over that file, it is what the agent
+// runs on. A change from what it runs on has the agent start again on it,
+// its engines serving throughout.
+message AgentSettings {
+  // yaml is agent.yaml's keys, as YAML; empty for none.
+  string yaml = 1;
+}
+
 // ControlPlanes is every control plane an agent may dial, sent by the leader
 // when an agent registers and whenever the list changes. The agent keeps it
 // beside its configuration, so a second control plane added after the node
@@ -266,6 +289,11 @@ message RunningShard {
   // restart could not be compared with the catalog — which is exactly when the
   // catalog is most likely to have been edited.
   string profile_fp = 15;
+
+  // compute is where its engine said it put the model when it became ready,
+  // kept by the agent across its own restarts: a control plane that takes
+  // over, or restarts, learns it here.
+  Compute compute = 16;
 }
 
 message GPU {
@@ -605,6 +633,18 @@ message ProvisionStatus {
   string detail = 5;
 }
 
+// Compute is where an engine runs its model, as it said when it loaded it.
+message Compute {
+  // backend is "CUDA", "Vulkan", "Metal" or "CPU".
+  string backend = 1;
+  // devices are the devices it named: "CUDA0: Quadro RTX 6000 (7.5)".
+  repeated string devices = 2;
+  // layers_offloaded of layers_total went to the devices; both 0 when the
+  // engine did not say.
+  int32 layers_offloaded = 3;
+  int32 layers_total = 4;
+}
+
 // ShardStatus reports a shard's lifecycle transition.
 message ShardStatus {
   string node_id = 1;
@@ -626,6 +666,10 @@ message ShardStatus {
   // going to look.
   FetchProgress fetch = 8;
 
+  // compute is what the engine says it computes on, read from its own output
+  // as it became ready: sent with STATE_READY when the engine says.
+  Compute compute = 9;
+
   enum State {
     STATE_UNSPECIFIED = 0;
     STATE_PENDING = 1;
@@ -844,3 +888,55 @@ message LoraAdapter {
   string name = 1;
   string source = 2;
 }
+
+// TrainStart has a node train a LoRA adapter (BACKLOG F4): on a base model's
+// weights from its registry, on question/answer pairs, with a Python trainer
+// the agent carries, in an environment it makes over an installed engine's.
+// One at a time per node.
+message TrainStart {
+  string job_id = 1;
+  // base_repo and base_revision name the base model's weights on its
+  // registry (Hugging Face): safetensors, not the GGUF a catalog may serve.
+  string base_repo = 2;
+  string base_revision = 3;
+  // data is the training pairs, NDJSON of {"q", "a"}.
+  bytes data = 4;
+  // params is the trainer's settings as JSON: rank, alpha, epochs, lr,
+  // batch, max_len.
+  string params = 5;
+  // gpu_memory_mib is what the job needs of a card; the agent starts it
+  // only when one has that much free, which lets it run beside an engine.
+  int32 gpu_memory_mib = 6;
+  // gguf asks for the adapter in llama.cpp's format too.
+  bool gguf = 7;
+}
+
+// TrainCancel stops a training job.
+message TrainCancel {
+  string job_id = 1;
+}
+
+// TrainStatus is how a training job is going, sent as it changes.
+message TrainStatus {
+  string job_id = 1;
+  // state is preparing, running, done or failed.
+  string state = 2;
+  int32 step = 3;
+  int32 steps = 4;
+  double loss = 5;
+  // message is what is being done, or why it failed.
+  string message = 6;
+  // peak_mib is the most the job held of the card.
+  int32 peak_mib = 7;
+}
+
+// TrainArtifact is a piece of what a training job made, sent before its
+// final status: a gzipped tar of the PEFT adapter (kind peft), or the GGUF
+// adapter (kind gguf).
+message TrainArtifact {
+  string job_id = 1;
+  string kind = 2;
+  int32 seq = 3;
+  bytes data = 4;
+  bool last = 5;
+}

cmd/llmgw-kb/main.go

@@ -0,0 +1,467 @@
+// Command llmgw-kb puts a person's documents into an llmgw knowledge base
+// from their own machine (BACKLOG F2): it reads the files there, cuts them
+// into chunks, has the gateway embed them with the person's key, and sends
+// the text and vectors — only for what changed since the last run, which a
+// ledger on the machine remembers.
+//
+// Its own binary, for Windows, macOS and Linux: it carries none of the
+// server, agent or setup code that llmgw does.
+package main
+
+import (
+	"bufio"
+	"context"
+	"encoding/json"
+	"errors"
+	"fmt"
+	"os"
+	"os/signal"
+	"path/filepath"
+	"strings"
+	"text/tabwriter"
+	"time"
+
+	"github.com/spf13/cobra"
+	"golang.org/x/term"
+
+	"gitea.mynodes.xyz/ilker/llmgw/internal/kb/extract"
+	"gitea.mynodes.xyz/ilker/llmgw/internal/kb/kbsync"
+	"gitea.mynodes.xyz/ilker/llmgw/internal/version"
+)
+
+func main() {
+	if err := root().Execute(); err != nil {
+		fmt.Fprintln(os.Stderr, "error:", err)
+		os.Exit(1)
+	}
+}
+
+// config is what `login` keeps: the gateway, the key, and for a personal key
+// the project it acts in.
+type config struct {
+	Gateway string `json:"gateway"`
+	Key     string `json:"key"`
+	Project string `json:"project,omitempty"`
+	CA      string `json:"ca,omitempty"`
+}
+
+func configPath() (string, error) {
+	dir, err := os.UserConfigDir()
+	if err != nil {
+		return "", err
+	}
+	return filepath.Join(dir, "llmgw-kb", "config.json"), nil
+}
+
+// loadConfig reads the saved configuration, with the environment over it:
+// LLMGW_URL, LLMGW_API_KEY, LLMGW_PROJECT.
+func loadConfig() (config, error) {
+	var c config
+	if p, err := configPath(); err == nil {
+		if b, err := os.ReadFile(p); err == nil {
+			if err := json.Unmarshal(b, &c); err != nil {
+				return c, fmt.Errorf("%s: %w", p, err)
+			}
+		}
+	}
+	for env, dst := range map[string]*string{"LLMGW_URL": &c.Gateway, "LLMGW_API_KEY": &c.Key, "LLMGW_PROJECT": &c.Project} {
+		if v := strings.TrimSpace(os.Getenv(env)); v != "" {
+			*dst = v
+		}
+	}
+	return c, nil
+}
+
+func client(flags *globalFlags) (*kbsync.Client, error) {
+	c, err := loadConfig()
+	if err != nil {
+		return nil, err
+	}
+	if flags.gateway != "" {
+		c.Gateway = flags.gateway
+	}
+	if flags.project != "" {
+		c.Project = flags.project
+	}
+	if flags.ca != "" {
+		c.CA = flags.ca
+	}
+	if c.Gateway == "" || c.Key == "" {
+		return nil, errors.New("no gateway or key: run `llmgw-kb login`, or set LLMGW_URL and LLMGW_API_KEY")
+	}
+	h, err := kbsync.NewHTTP(c.CA)
+	if err != nil {
+		return nil, err
+	}
+	return &kbsync.Client{Base: c.Gateway, Key: c.Key, Project: c.Project, HTTP: h}, nil
+}
+
+type globalFlags struct {
+	gateway, project, ca, ledger string
+}
+
+func (g *globalFlags) openLedger(ctx context.Context) (*kbsync.Ledger, error) {
+	p := g.ledger
+	if p == "" {
+		var err error
+		if p, err = kbsync.DefaultLedgerPath(); err != nil {
+			return nil, err
+		}
+	}
+	return kbsync.OpenLedger(ctx, p)
+}
+
+func root() *cobra.Command {
+	g := &globalFlags{}
+	cmd := &cobra.Command{
+		Use:   "llmgw-kb",
+		Short: "Put your documents into an llmgw knowledge base, from this machine",
+		Long: `llmgw-kb reads documents on this machine, cuts them into chunks, has the
+gateway embed them with your key, and sends the text and vectors to a
+knowledge base. Files are never uploaded. Run it as often as you like: a
+ledger on this machine remembers what was sent, and only what changed is
+read and sent again.`,
+		Version:       version.Version,
+		SilenceUsage:  true,
+		SilenceErrors: true,
+	}
+	cmd.PersistentFlags().StringVar(&g.gateway, "gateway", "", "the gateway's address (default: from login, or LLMGW_URL)")
+	cmd.PersistentFlags().StringVar(&g.project, "project", "", "the project a personal key acts in (default: from login, or LLMGW_PROJECT)")
+	cmd.PersistentFlags().StringVar(&g.ca, "ca", "", "a PEM file of certificates to trust besides the system's")
+	cmd.PersistentFlags().StringVar(&g.ledger, "ledger", "", "the ledger file (default: in your configuration directory)")
+	cmd.AddCommand(loginCmd(g), collectionsCmd(g), syncCmd(g), statusCmd(g), historyCmd(g), formatsCmd())
+	return cmd
+}
+
+func loginCmd(g *globalFlags) *cobra.Command {
+	return &cobra.Command{
+		Use:   "login",
+		Short: "Save the gateway's address and your key",
+		Long: `Asks for the gateway's address and your API key, checks them, and saves them
+in your configuration directory, readable by you only. The key is typed
+without being shown, so it does not end up in your shell's history.`,
+		Args: cobra.NoArgs,
+		RunE: func(cmd *cobra.Command, _ []string) error {
+			c, _ := loadConfig()
+			in := bufio.NewReader(os.Stdin)
+			ask := func(q, def string) string {
+				if def != "" {
+					fmt.Fprintf(cmd.OutOrStdout(), "%s [%s]: ", q, def)
+				} else {
+					fmt.Fprintf(cmd.OutOrStdout(), "%s: ", q)
+				}
+				s, _ := in.ReadString('\n')
+				if s = strings.TrimSpace(s); s == "" {
+					return def
+				}
+				return s
+			}
+			c.Gateway = strings.TrimRight(ask("Gateway address (https://…)", or(g.gateway, c.Gateway)), "/")
+			fmt.Fprint(cmd.OutOrStdout(), "API key: ")
+			var key string
+			if term.IsTerminal(int(os.Stdin.Fd())) {
+				b, err := term.ReadPassword(int(os.Stdin.Fd()))
+				fmt.Fprintln(cmd.OutOrStdout())
+				if err != nil {
+					return err
+				}
+				key = strings.TrimSpace(string(b))
+			} else {
+				key, _ = in.ReadString('\n')
+				key = strings.TrimSpace(key)
+			}
+			if key != "" {
+				c.Key = key
+			}
+			c.Project = ask("Project, for a personal key (empty for a project's key)", or(g.project, c.Project))
+			if g.ca != "" {
+				c.CA = g.ca
+			}
+			h, err := kbsync.NewHTTP(c.CA)
+			if err != nil {
+				return err
+			}
+			cl := &kbsync.Client{Base: c.Gateway, Key: c.Key, Project: c.Project, HTTP: h, Wait: time.Second}
+			ctx, cancel := context.WithTimeout(cmd.Context(), 30*time.Second)
+			defer cancel()
+			cs, err := cl.Collections(ctx)
+			if err != nil {
+				return fmt.Errorf("the gateway refused: %w", err)
+			}
+			p, err := configPath()
+			if err != nil {
+				return err
+			}
+			if err := os.MkdirAll(filepath.Dir(p), 0o700); err != nil {
+				return err
+			}
+			b, _ := json.MarshalIndent(c, "", "  ")
+			if err := os.WriteFile(p, b, 0o600); err != nil {
+				return err
+			}
+			fmt.Fprintf(cmd.OutOrStdout(), "Saved in %s. This key reaches %d collection(s); `llmgw-kb collections` lists them.\n", p, len(cs))
+			return nil
+		},
+	}
+}
+
+func or(a, b string) string {
+	if a != "" {
+		return a
+	}
+	return b
+}
+
+func collectionsCmd(g *globalFlags) *cobra.Command {
+	return &cobra.Command{
+		Use:   "collections",
+		Short: "List the collections your key may read, and which it may add to",
+		Args:  cobra.NoArgs,
+		RunE: func(cmd *cobra.Command, _ []string) error {
+			c, err := client(g)
+			if err != nil {
+				return err
+			}
+			cs, err := c.Collections(cmd.Context())
+			if err != nil {
+				return err
+			}
+			if len(cs) == 0 {
+				fmt.Fprintln(cmd.OutOrStdout(), "No collections. An administrator creates them and grants access (llmgw ctl kb).")
+				return nil
+			}
+			tw := tabwriter.NewWriter(cmd.OutOrStdout(), 2, 4, 2, ' ', 0)
+			fmt.Fprintln(tw, "NAME\tACCESS\tOWNER\tEMBEDDED BY")
+			for _, x := range cs {
+				access := "read"
+				if x.Writable {
+					access = "read, write"
+				}
+				fmt.Fprintf(tw, "%s\t%s\t%s\t%s (%d)\n", x.Name, access, or(x.OwnerName, x.Owner), x.EmbedModel, x.Dims)
+			}
+			return tw.Flush()
+		},
+	}
+}
+
+func syncCmd(g *globalFlags) *cobra.Command {
+	var o kbsync.Options
+	var verbose bool
+	var ocr string
+	var maxMiB int64
+	cmd := &cobra.Command{
+		Use:   "sync COLLECTION PATH...",
+		Short: "Send a collection what changed in these folders and files",
+		Long: `Reads the files under each PATH that the collection does not already have as
+they are, and sends their text and vectors. Run it again whenever you like;
+unchanged files are not even read.
+
+Documents are named by their path under PATH, with forward slashes; with
+several PATHs, each folder's own name comes first.
+
+  --dry-run   reads and cuts everything, sends nothing, and says what would
+              leave this machine
+  --delete    removes documents this machine sent whose files are gone
+              (only under these PATHs; nothing anybody else sent)
+  --retry     reads again files skipped before (after installing LibreOffice
+              or Tesseract, say)
+
+Without --include, every file of a known format is read (llmgw-kb formats);
+hidden files, version control folders and Office lock files are not.`,
+		Args: cobra.MinimumNArgs(2),
+		RunE: func(cmd *cobra.Command, args []string) error {
+			c, err := client(g)
+			if err != nil {
+				return err
+			}
+			ctx, stop := signal.NotifyContext(cmd.Context(), os.Interrupt)
+			defer stop()
+			l, err := g.openLedger(ctx)
+			if err != nil {
+				return err
+			}
+			defer l.Close()
+			o.Collection, o.Roots = args[0], args[1:]
+			o.Extract = extract.Options{Helpers: extract.FindHelpers(), OCRLanguages: ocr, MaxBytes: maxMiB << 20}
+			out := cmd.OutOrStdout()
+			c.Waiting = func(why string, d time.Duration) {
+				fmt.Fprintf(out, "  ..       waiting %s: %s\n", d, why)
+			}
+			start := time.Now()
+			rep, err := kbsync.Sync(ctx, c, l, o, func(e kbsync.Event) {
+				switch e.Kind {
+				case "unchanged":
+					if !verbose {
+						return
+					}
+					fmt.Fprintf(out, "  =        %s\n", e.Name)
+				case "sent", "would-send":
+					fmt.Fprintf(out, "  %-8s %s (%d chunks, %s)\n", label(e.Kind), e.Name, e.Chunks, size(e.Bytes))
+				default:
+					line := fmt.Sprintf("  %-8s %s", label(e.Kind), e.Name)
+					if e.Detail != "" {
+						line += ": " + e.Detail
+					}
+					fmt.Fprintln(out, line)
+				}
+			})
+			if rep.Collection.Name != "" {
+				verb := "sent"
+				if o.DryRun {
+					verb = "would send"
+				}
+				fmt.Fprintf(out, "\n%s: %d files; %s %d (%d chunks, %s), %d unchanged, %d skipped, %d failed, %d deleted, in %s\n",
+					rep.Collection.Name, rep.Files, verb, rep.Sent, rep.Chunks, size(rep.Bytes), rep.Unchanged,
+					rep.Skipped, rep.Failed, rep.Deleted, time.Since(start).Round(time.Second))
+			}
+			if err != nil {
+				return err
+			}
+			if rep.Failed > 0 {
+				return fmt.Errorf("%d document(s) failed; run again to retry them", rep.Failed)
+			}
+			return nil
+		},
+	}
+	f := cmd.Flags()
+	f.StringSliceVar(&o.Include, "include", nil, "read only files matching these globs (\"*.pdf,docs/**/*.md\")")
+	f.StringSliceVar(&o.Exclude, "exclude", nil, "never read files matching these globs")
+	f.BoolVar(&o.Hidden, "hidden", false, "read hidden files and folders too")
+	f.BoolVar(&o.DryRun, "dry-run", false, "send nothing; say what would be sent")
+	f.BoolVar(&o.Delete, "delete", false, "remove documents this machine sent whose files are gone")
+	f.BoolVar(&o.Retry, "retry", false, "read again files skipped before")
+	f.IntVar(&o.Workers, "workers", 4, "files read and sent at once")
+	f.StringVar(&ocr, "ocr-languages", "tur+eng", "Tesseract's languages for images and scanned pages")
+	f.Int64Var(&maxMiB, "max-size", 256, "skip files larger than this many MiB")
+	f.BoolVarP(&verbose, "verbose", "v", false, "list unchanged files too")
+	return cmd
+}
+
+func label(kind string) string {
+	switch kind {
+	case "would-send":
+		return "send"
+	case "would-delete":
+		return "delete"
+	}
+	return kind
+}
+
+func size(b int64) string {
+	switch {
+	case b >= 1<<20:
+		return fmt.Sprintf("%.1f MiB", float64(b)/(1<<20))
+	case b >= 1<<10:
+		return fmt.Sprintf("%.0f KiB", float64(b)/(1<<10))
+	}
+	return fmt.Sprintf("%d B", b)
+}
+
+func statusCmd(g *globalFlags) *cobra.Command {
+	var all bool
+	return &cobra.Command{
+		Use:   "status COLLECTION",
+		Short: "Show what this machine has sent to a collection, and what it skipped",
+		Args:  cobra.ExactArgs(1),
+		RunE: func(cmd *cobra.Command, args []string) error {
+			c, err := client(g)
+			if err != nil {
+				return err
+			}
+			m, err := c.Manifest(cmd.Context(), args[0])
+			if err != nil {
+				return err
+			}
+			l, err := g.openLedger(cmd.Context())
+			if err != nil {
+				return err
+			}
+			defer l.Close()
+			es, err := l.Entries(cmd.Context(), strings.TrimRight(c.Base, "/"), m.Collection.ID)
+			if err != nil {
+				return err
+			}
+			counts := map[kbsync.State]int{}
+			tw := tabwriter.NewWriter(cmd.OutOrStdout(), 2, 4, 2, ' ', 0)
+			fmt.Fprintln(tw, "STATE\tDOCUMENT\tCHUNKS\tWHEN\tWHY")
+			for _, e := range es {
+				counts[e.State]++
+				if e.State == kbsync.Sent && !all {
+					continue
+				}
+				fmt.Fprintf(tw, "%s\t%s\t%d\t%s\t%s\n", e.State, e.Name, e.Chunks, e.At.Format("2006-01-02 15:04"), e.Reason)
+			}
+			_ = tw.Flush()
+			fmt.Fprintf(cmd.OutOrStdout(), "\n%s holds %d documents; from this machine %d sent, %d skipped, %d failed.",
+				m.Collection.Name, len(m.Documents), counts[kbsync.Sent], counts[kbsync.Skipped], counts[kbsync.Failed])
+			if !all {
+				fmt.Fprint(cmd.OutOrStdout(), " (--all lists the sent ones too)")
+			}
+			fmt.Fprintln(cmd.OutOrStdout())
+			return nil
+		},
+	}
+}
+
+func historyCmd(g *globalFlags) *cobra.Command {
+	return &cobra.Command{
+		Use:   "history",
+		Short: "List the latest syncs from this machine",
+		Args:  cobra.NoArgs,
+		RunE: func(cmd *cobra.Command, _ []string) error {
+			l, err := g.openLedger(cmd.Context())
+			if err != nil {
+				return err
+			}
+			defer l.Close()
+			runs, err := l.Runs(cmd.Context(), 20)
+			if err != nil {
+				return err
+			}
+			tw := tabwriter.NewWriter(cmd.OutOrStdout(), 2, 4, 2, ' ', 0)
+			fmt.Fprintln(tw, "STARTED\tTOOK\tCOLLECTION\tSENT\tUNCHANGED\tSKIPPED\tFAILED\tDELETED")
+			for _, r := range runs {
+				took := "unfinished"
+				if !r.Finished.IsZero() {
+					took = r.Finished.Sub(r.Started).Round(time.Second).String()
+				}
+				fmt.Fprintf(tw, "%s\t%s\t%s\t%d\t%d\t%d\t%d\t%d\n", r.Started.Format("2006-01-02 15:04"), took,
+					r.Collection, r.Sent, r.Unchanged, r.Skipped, r.Failed, r.Delete)
+			}
+			return tw.Flush()
+		},
+	}
+}
+
+func formatsCmd() *cobra.Command {
+	return &cobra.Command{
+		Use:   "formats",
+		Short: "List the formats read, and the helpers found on this machine",
+		Args:  cobra.NoArgs,
+		RunE: func(cmd *cobra.Command, _ []string) error {
+			builtIn, helper := extract.Supported()
+			out := cmd.OutOrStdout()
+			fmt.Fprintf(out, "Read here:\n%s\n\n", wrap(builtIn))
+			fmt.Fprintf(out, "Read with a helper program:\n%s\n\n", wrap(helper))
+			fmt.Fprintln(out, "Helpers:")
+			for _, l := range extract.FindHelpers().Describe() {
+				fmt.Fprintln(out, "  "+l)
+			}
+			return nil
+		},
+	}
+}
+
+func wrap(words []string) string {
+	var b strings.Builder
+	line := "  "
+	for _, w := range words {
+		if len(line)+len(w) > 78 {
+			b.WriteString(strings.TrimRight(line, " ") + "\n")
+			line = "  "
+		}
+		line += w + " "
+	}
+	b.WriteString(strings.TrimRight(line, " "))
+	return b.String()
+}

cmd/llmgw-kb/main_test.go

@@ -0,0 +1,24 @@
+package main
+
+import (
+	"os/exec"
+	"strings"
+	"testing"
+)
+
+// The client goes onto employees' laptops: none of the server, agent, store
+// or setup code may come with it, and it must build where llmgw does not.
+func TestTheClientCarriesNoServerCode(t *testing.T) {
+	out, err := exec.Command("go", "list", "-deps", ".").Output()
+	if err != nil {
+		t.Skipf("go list: %v", err)
+	}
+	for _, dep := range strings.Fields(string(out)) {
+		for _, banned := range []string{"/internal/store", "/internal/router", "/internal/agent", "/internal/control",
+			"/internal/cli", "/internal/guard", "github.com/jackc/pgx"} {
+			if strings.Contains(dep, banned) {
+				t.Errorf("llmgw-kb depends on %s", dep)
+			}
+		}
+	}
+}

docs.go

@@ -39,7 +39,7 @@ func (d Doc) Source() string {
 // than with either reader, because the console's manual and the gateway's
 // docs portal both list them, and two lists drift.
 var docOrder = []string{"console", "quick-install", "install", "node-onboarding", "onboarding",
-	"clients", "guardrails", "choosing-models", "fabrics", "facility-aware-placement", "data",
+	"clients", "guardrails", "knowledge", "choosing-models", "fabrics", "facility-aware-placement", "data",
 	"reference", "development"}
 
 // Manual is every document, in reading order: those docOrder names, then any

docs/admin-api.md

@@ -8,8 +8,9 @@ description is.
 anything they do can be done with `curl`, a generated client or the
 [Terraform provider](../terraform-provider-llmgw/README.md). It adds and
 removes models, scales them, cordons and upgrades nodes, sets capacity
-quotas, attaches guardrail policies, and reads the fleet, usage and the audit
-log.
+quotas, attaches guardrail policies, adds organizations, teams and projects,
+issues and revokes API keys, puts people in the tree, and reads the fleet,
+usage and the audit log.
 
 ## Reaching it
 
@@ -99,6 +100,14 @@ curl -s "${H[@]}" -X POST $API/models/qwen-small/scale -d '{"min": 0, "max": 2,
 # A team's capacity quota.
 curl -s "${H[@]}" -X PUT $API/quotas -d '{"scope": "team:Acme/ml", "limit": {"gpus": 8}}'
 
+# An organization, a team in it and a project in that; then a key for the
+# project. The answer holds the key once: only its hash is kept.
+curl -s "${H[@]}" -X POST $API/orgs -d '{"name": "Acme"}'
+curl -s "${H[@]}" -X POST $API/teams -d '{"name": "ml", "in": "Acme"}'
+curl -s "${H[@]}" -X POST $API/projects -d '{"name": "search", "in": "Acme/ml"}'
+curl -s "${H[@]}" -X POST $API/keys -d '{"name": "ci", "project": "Acme/ml/search", "expires_in": "720h"}'
+curl -s "${H[@]}" -X POST $API/users/ayse@example.com/memberships -d '{"target": "team:Acme/ml"}'
+
 # Remove the model; a DELETE carries the content type too.
 curl -s "${H[@]}" -X DELETE $API/models/qwen-small
 ```
@@ -149,11 +158,11 @@ go generate ./internal/docgen
 
 ## What it does not cover
 
-- **Organizations, teams, projects, people and API keys.** These live in the
-  store and are managed with `llmgw org`, `llmgw users` and `llmgw keys` on the
-  control plane, or in the web console's `/admin` pages on the client port
-  behind a session. The admin API reads the tree only where a quota or a
-  policy names a part of it.
+- **Making people's accounts.** People sign up, are invited, or come from
+  the identity provider (OIDC, SCIM); the API lists them and puts them in
+  the tree, and `llmgw users` approves, disables and deletes them. Key
+  lifetime limits and spending limits are `llmgw keys ttl` and `llmgw keys
+  limits`.
 - **The client-facing API** (`/v1/chat/completions` and the rest), which is
   OpenAI's and Anthropic's, on `listen`, with API keys.
 - **The rest of the admin listener**: `/metrics`, `/health`, the dashboard

docs/console.md

@@ -164,6 +164,15 @@ which): requests, tokens, errors and spend.
 team or project is guaranteed and may at most hold, and what it holds now.
 **A** adds one; on a row, change or remove it.
 
+**Knowledge** lists the knowledge-base collections: each one's owner, the
+model that embeds it, how many documents and chunks it holds and who else it
+is granted to. **A** makes one. On a row: its **Documents**, **Grant…** a
+project, team, organization, person or key read or write access, **Revoke**
+a grant, **Answer from it by default…** for a part of the tree, **Try a
+search…** as an administrator, or **Delete…** it. People
+fill collections from their own machines with `llmgw-kb`
+([knowledge.md](knowledge.md)). Off on a SQLite store, and says why.
+
 **Mirror** is what the model mirror holds, one row for each version of each
 model — a file, or the set of files a model of several is: shards, a vision
 projector, a config and tokenizer — with how far its download has got and

docs/guardrails.md

@@ -4,7 +4,7 @@ Written for the operator deciding what the gateway lets through: what a
 policy can check, what it can do, and how one is attached to a team.
 
 `llmgw server init` writes a `baseline` policy that applies to every
-request, with Llama Guard 3 1B as its judge on a CPU; see
+request, with Qwen3Guard-Gen 0.6B as its judge on CPUs; see
 [install.md](install.md#2-the-control-plane) for what it checks. The
 policies are part of the cluster's settings, kept in the store: `llmgw ctl
 config edit` (or the console's Settings) changes them, and the change is in
@@ -15,6 +15,19 @@ checked and nothing costs anything. With one, every request is judged once it is
 within its limits, with its stored conversation spliced in, and before it is
 sent to a model; and the answer is judged before the client sees it.
 
+## Turning them off and on
+
+```
+llmgw ctl guard off --note "why"   # no request is checked, whatever the policies say
+llmgw ctl guard on                 # every policy checks again, as it was
+```
+
+That is `guard.enabled` in the cluster's settings, in force on every control
+plane within seconds without a restart. The policies and where they are
+attached are kept. While off, `llmgw ctl status`, `llmgw ctl guard list` and
+the guardrails page say so. To stop one policy rather than all, give it
+`mode: off`, or `mode: shadow` to have it record what it would do.
+
 ## A policy
 
 ```yaml
@@ -101,44 +114,136 @@ guard:
 
 ## A judge model
 
-Patterns find what has a shape. For what does not — violence, self-harm,
-abuse, whatever a moderation model was trained on — a policy can have a
-**judge model on the cluster** read the text:
+Patterns find what has a shape. For what does not — violence, crime,
+sexual content, self-harm, jailbreaks, whatever a moderation model was
+trained on — a policy can have a **judge model on the cluster** read the
+text.
+
+The judge must be in the catalog under the name the policy gives, as a
+judge (`role: guard`), with a replica ready. When it is not, what the
+policy judges is refused with `guard_unavailable` (`fail: closed`) or served
+unjudged (`fail: open`), and `llmgw ctl status` and the console's Overview
+say which judge and why.
+
+### Which judge
+
+| judge | runs on | size | |
+|---|---|---|---|
+| `qwen3guard-0.6b` | CPUs | 0.8 GB | the default: about 2 s a check on two threads |
+| `qwen3guard-4b` | a GPU | 2.5 GB | for a cluster with a card to spare |
+| `qwen3guard-8b` | a GPU | 4.7 GB | |
+| `llama-guard-3-1b` | CPUs | 0.9 GB | the earlier default: refuses ordinary coding requests |
+| `llama-guard-3-8b` | a GPU | 4.6 GB | |
+
+Measured on the reference cluster, as the judge's probability of "unsafe"
+for ordinary coding requests against harmful ones, Llama Guard 3 1B could
+not tell them apart: "continue" 0.51, "continue the code above" 0.89, "how
+do I kill all child processes in bash" 0.92, against 0.97 to 1.00 for
+harmful requests, at any quantisation. Qwen3Guard-Gen 0.6B gave the
+ordinary ones 0.15 at the most and the harmful ones 1.00. Its verdicts are
+safe, unsafe and **controversial**; controversial counts as safe unless the
+policy says `controversial: unsafe`.
+
+A policy that names one of these is enough: the leader fetches its weights
+into the mirror and serves it as a judge, and serves it again if it is
+removed from the catalog. Another judge is served by hand with `--repo`.
+
+```
+llmgw ctl guard judge status              # each judge the policies ask, and what is wrong with it
+llmgw ctl guard judge enable              # Qwen3Guard-Gen 0.6B on CPUs, or the judge the policies ask that is missing
+llmgw ctl guard judge enable --gpu        # Qwen3Guard-Gen 4B on a GPU
+llmgw ctl guard judge use qwen3guard-4b qwen3guard-0.6b   # the order every policy asks them in
+llmgw ctl guard judge enable --replace    # when a model of that name is there, but not as a judge
+```
+
+or **Do → Serve the guardrails' judge** in the console.
+
+### A GPU and CPUs together
+
+`judge.models` is a list asked in turn: the first judge that answers gives
+the verdict. The next is asked when one has no replica serving, fails, or
+does not answer within `latency_budget`; a judge with no replica at all is
+passed over at once. A judge on a GPU first and one on CPUs after it is
+the GPU's speed while it has its card, and the CPUs' when it does not.
+
+```yaml
+judge:
+  models: [qwen3guard-4b, qwen3guard-0.6b]
+  input: block
+  output: block
+  latency_budget: 5s           # each judge's wait
+```
+
+A judge on a GPU is served as a `standard` model, on one card: it takes a
+card only from what may give one up, and waits while every card is busy,
+the judge on CPUs standing in. A judge on CPUs is `critical`, with two
+replicas, because every judged request waits on one. `llmgw ctl guard judge
+status` says, for each judge, which is asked when it gives no verdict.
+
+### When it refuses ordinary requests
+
+A refusal by a judge names the hazard it saw — the client is told `refused
+by the gateway's policy (judge: Violent)` — and the same is kept with the
+request: `llmgw ctl request <id>`, the server's log (`guard_rule`) and the
+guardrails page show it. To see what the judge says of any text, without
+anything being kept:
+
+```
+llmgw ctl guard judge ask 'How do I sort a list in Python?'
+# qwen3guard-0.6b: safe, 1900 ms
+# it answered: "Safety: Safe\nCategories: None"
+```
+
+A judge that says `unsafe` to ordinary text is usually Llama Guard 3 1B
+(`llmgw ctl guard judge use qwen3guard-0.6b`), the wrong weights (another
+repository's, without the judge's template: `judge enable --replace` serves
+the right ones), or a policy counting every category: give its judge
+`categories`. The baseline counts Violent, Non-violent Illegal Acts, Sexual
+Content or Sexual Acts, Suicide & Self-Harm and Jailbreak; PII is the
+patterns', and copyright, unethical acts and politically sensitive topics
+refuse too much ordinary work. While finding out, `judge.input: flag`
+records instead of refusing.
+
+A judge's file served from the mirror page — Llama Guard, Qwen3Guard,
+ShieldGemma — is served as a judge too, under the name the policies ask.
+By hand, the catalog entry is:
 
 ```yaml
 # a catalog model (server.yaml models:, or ctl model add)
-- name: llama-guard-3-1b
-  source: http://mirror:8081/llama-guard-3-1b/…/Llama-Guard-3-1B-Q4_K_M.gguf
+- name: qwen3guard-0.6b
+  source: http://mirror:8081/qwen3guard-0.6b/…/Qwen3Guard-Gen-0.6B.Q8_0.gguf
   role: guard
   guard:
-    format: llamaguard3        # llamaguard3, shieldgemma or json
-    categories: [Violent Crimes, Non-Violent Crimes, Sex-Related Crimes, …]  # names for S1, S2, …
+    format: qwen3guard         # qwen3guard, llamaguard3, shieldgemma or json
   scaling: {min: 2, max: 2}
   profiles: [{name: cpu, engine: llamacpp}]
 
 # in a guard policy
 judge:
-  model: llama-guard-3-1b
+  models: [qwen3guard-0.6b]    # or model: one judge
   input: block                 # flag or block; empty does not ask about input
   output: flag
-  categories: [S1, S10]        # only these count; empty: any unsafe verdict
-  latency_budget: 250ms        # then the policy's fail: open serves (flagged), closed refuses
+  categories: [Violent, Jailbreak]   # only these count; empty: any unsafe verdict
+  controversial: safe          # or unsafe
+  latency_budget: 250ms        # each judge's wait; after the last, fail: open serves (flagged), closed refuses
 ```
 
 A model with `role: guard` is the guardrails' alone:
 
 - it is not in `/v1/models` or the chat page, and is refused (404) to
   every key, as a name and as a fallback;
-- it is `critical` by default and must always run — on demand is refused.
-  Give it two replicas on different nodes, because every judged request
-  waits on it.
+- it must always run — on demand is refused. One on CPUs is `critical`
+  and should have two replicas on different nodes, because every judged
+  request waits on it.
 
 The gateway's calls to it go past admission, accounting and the guardrails
 themselves.
 
 - **Formats:**
-  - `llamaguard3` reads Meta's "safe" / "unsafe
-S1,S10".
+  - `qwen3guard` asks in Qwen3Guard's own words and reads "Safety: Safe |
+    Unsafe | Controversial" and its "Categories:".
+  - `llamaguard3` reads Meta's "safe" / "unsafe\nS1,S10"; give the catalog
+    entry `categories` to have S1 recorded as its name.
   - `shieldgemma` asks Google's model whether the text breaks the listed
     principles (`categories`, each a sentence) and reads Yes or No.
   - `json` expects `{"safe": false, "categories": [...]}`.
@@ -148,6 +253,9 @@ S1,S10".
     is read every 600 characters and at the end, so text between two
     readings has gone out already; `stream: buffer_all` holds it until it
     is read.
+  - **A long text** is read by its start and its end (about 6000
+    characters), so a message with a file pasted into it fits the judge's
+    context rather than overflowing it and going unjudged.
 - **Cache:** the same text is not asked about twice within five minutes.
 
 ## Where a policy applies

docs/install.md

@@ -129,8 +129,9 @@ That is the whole step. In order, it:
    `nvidia-smi` finds any, on CPU otherwise — and waits for it to install its
    engine
 9. closes llmgw's internal ports to everything but the cluster's networks
-10. fetches Llama Guard 3 1B, the guardrails' judge, and runs two replicas
-    on the CPUs of nodes with room — this one, or the first that join
+10. fetches Qwen3Guard-Gen 0.6B, the guardrails' judge, and runs two
+    replicas on the CPUs of nodes with room — this one, or the first that
+    join ([guardrails.md](guardrails.md#a-judge-model) has one on a GPU too)
 11. fetches Qwen2.5 0.5B into the mirror and serves it on this node, so the
     first thing anyone opens `/chat` to find is a model that answers. A
     machine without room for it is told so within two minutes, and why.
@@ -200,6 +201,16 @@ this by running the same command again, on the control plane that holds the
 authority: etcd is rebuilt on TLS with Patroni in maintenance mode, so
 PostgreSQL serves throughout and no failover happens.
 
+Each member's PostgreSQL is sized to the machine's memory, read each time
+the command runs: a quarter of it as cache (`shared_buffers`), three quarters
+counted on for the planner, and room to sort and build indexes, all capped. A
+member that also runs a control plane or an agent gets less: about 15% as
+cache, so the gateway is not pushed into swap. A knowledge base's vector
+index is searched well only from memory, so give a store machine more RAM and
+run the command again. The new cache takes a PostgreSQL restart, which Patroni
+does one member at a time, the leader last. The leader's restart costs a few
+seconds without writes, and there is no failover.
+
 Every step looks before it acts. Running it again finishes an install that
 stopped halfway and changes nothing on one that did not; it never overwrites
 `server.yaml`, the authority, or an existing account's password.
@@ -257,8 +268,14 @@ llmgw ctl config history                 # who changed what, when, and why
 llmgw ctl config rollback 4              # back to version 4, as a new version
 ```
 
-or on the console's **Settings** screen. A change is checked before it is
-kept — one that would not load is refused — and recorded in the audit log.
+or on the console's **Settings** screen, or in a browser on the portal's
+**Settings** page (`/admin/settings`, for administrators only: the document
+can hold webhook secrets). The page edits the same document, lists its
+history, and puts an earlier version back. A change is checked before it is
+kept, so one that would not load is refused. A change made from a version
+someone else has replaced since is refused too, rather than overwriting
+theirs. Every change is recorded in the audit log, with the document's size
+but not its contents.
 Guardrails change live; anything else has every control plane start again on
 the new settings, the standbys first and the leader half a minute later, so
 one of them is always serving. Engines keep running through it.
@@ -268,6 +285,33 @@ cluster keys — with a `baseline` guard policy, if the file had none — and
 from then on those keys in any server.yaml are not read (the server says so
 when it starts).
 
+**The nodes' agents** are set the same way, under `agents`. `agents.default`
+is for every node, and `agents.nodes.<node-id>` is for one node, laid over
+the default:
+
+```
+llmgw ctl config set agents.default.heartbeat_interval 5s
+llmgw ctl config set agents.nodes.cuda01.labels.gpu rtx6000
+llmgw ctl config edit                    # fabrics, engines, supervisor, discovery…
+```
+
+Each node's `agent.yaml` then needs only what its agent must know before it
+can be told anything: `node_id`, `server` or `servers`, `hostname`,
+`mgmt_addr` and `tls`. Those keys are refused under `agents`.
+
+- **Precedence.** Anything else in the file is still read, with the
+  cluster's settings over it.
+- **Applying a change.** The control plane sends each node its settings, a
+  second apart. An agent whose settings changed keeps them beside its
+  `agent.yaml` (in `/var/lib/llmgw-agent/agent-settings.yaml`) and starts
+  again on them, its engines serving throughout.
+- **Bad settings.** Settings that would not load are refused when they are
+  saved. If they somehow do not load beside one node's own file, that agent
+  keeps running on what it has and logs why.
+- **Editing `agent.yaml` itself.** The agent notices the change within ten
+  seconds and starts again on it. If the edited file does not load, the
+  agent reports it and keeps running on what it had.
+
 **This machine's own keys** stay in `/etc/llmgw/server.yaml`: `listen`,
 `admin_listen`, `metrics_listen`, `agent_listen`, `agent_tls`, `auth.store`,
 `secrets`, `tracing`, `monitoring`, `mirror.api` and `mirror.tls`, the backup
@@ -1226,6 +1270,51 @@ prices and budgets, rate limits, usage records, and key allowlists.
 whose allowlist leaves the external model out cannot reach it, as a fallback or
 by name. The cluster page and `ctl models` mark it `external`.
 
+### Pictures from ComfyUI
+
+A ComfyUI server is an external model with `api: comfyui`. Clients ask it for
+pictures at OpenAI's `/v1/images/generations`; llmgw fills a ComfyUI workflow
+with the request, queues it, waits, and sends the pictures back as `b64_json`:
+
+```yaml
+- name: flux-schnell
+  source: external
+  external:
+    url: http://10.0.0.147:8188                  # ComfyUI, started with --listen
+    model: flux1-schnell-fp8.safetensors         # the checkpoint, as ComfyUI lists it
+    api: comfyui
+    workflow: flux-schnell                       # or flux-dev, or a JSON file
+```
+
+`flux-schnell` (the default) and `flux-dev` are built in. They suit the
+single-file FLUX checkpoints that ComfyUI loads with `CheckpointLoaderSimple`.
+For any other model, export the workflow from ComfyUI with **Save (API)** and
+name the file. In the file, put placeholders where the request's values go:
+
+- `{{prompt}}` and `{{negative_prompt}}` are the prompts.
+- `{{width}}` and `{{height}}` are the size.
+- `{{n}}` is the batch size.
+- `{{seed}}` is the seed.
+- `{{steps|4}}` and `{{guidance|3.5}}` take the value after the `|` when the
+  request leaves them out.
+- `{{checkpoint}}` is the `model` above.
+
+A request is OpenAI's: `prompt`, `n` (at most 4), `size` (each side a multiple
+of 16, from 256 to 2048) and `response_format: b64_json`. Beyond those, it may
+send `seed`, `steps`, `negative_prompt` and `guidance`. The seed used comes
+back in `x-llmgw-seed`. `url` is not offered, because pictures are kept
+nowhere. ComfyUI also writes the workflow into each PNG; llmgw strips it
+before the picture is sent.
+
+Requests for pictures pass the same checks as any other request: keys,
+allowlists, limits, admission, and the guardrails on the prompt. A client that
+leaves takes its workflow off the ComfyUI queue. Usage is recorded, but for
+now it is charged as the prompt's tokens; a per-picture price is not yet
+available.
+
+Start ComfyUI with `--disable-smart-memory` on a card that llmgw also places
+models on. Without it, ComfyUI keeps the model on the card after each picture.
+
 ### Federating clusters
 
 Another llmgw cluster — another site, another team's GPUs — can be a provider
@@ -1702,9 +1791,11 @@ the file again.
 
 On a machine that also runs the control plane, PostgreSQL's defaults assume it
 owns the box. A gateway swapped out to give the database cache is a worse
-outcome than a smaller cache, so cap it: `shared_buffers = 96MB` and
-`max_connections = 40` are enough for a store whose working set is measured in
-megabytes.
+outcome than a smaller cache, so size it to the machine. `server init` does this
+for a PostgreSQL it installs: `shared_buffers` about 15% of memory (at least
+96MB), `effective_cache_size` half, `work_mem` and `maintenance_work_mem`
+scaled to it, and `max_connections = 40`. A store with a large knowledge base
+wants a machine of its own, where its vector index fits in memory.
 
 ## Tagging requests
 
@@ -2150,13 +2241,29 @@ llmgw ctl config set firewall.enabled false   # every machine's table taken out
 llmgw ctl config set firewall.enabled true    # and put back as it was
 ```
 
-It is in force within seconds: each node's agent takes its table out or puts
-it back, and each control plane's `llmgw-firewall-sync.path` (installed by
-`server init` and `server add`; `sudo llmgw firewall sync --install` on one
-set up before) does the same there. The rules files stay, so turning it back
-on restores them. Nodes added while it is off get none. A separate mirror
-machine and the store's members keep theirs for now. `firewall_networks`
-changes reach the nodes the same way.
+It is in force on every machine within a minute, each in its own way:
+
+- **nodes**: the agent takes its table out or puts it back;
+- **control planes**: `llmgw-firewall-sync.path`, installed by `server
+  init` and `server add` (`sudo llmgw firewall sync --install` on one set up
+  before);
+- **a separate mirror machine**: the leading control plane tells the mirror,
+  which writes it to `.firewall.json` in its directory, and the same path
+  unit applies it. `mirror init` installs it; on a mirror set up before, run
+  `sudo llmgw mirror init` again with the same flags;
+- **store members that are nothing else**: they run no llmgw, so
+  `llmgw-firewall-follow.timer` reads what the leading control plane keeps in
+  the store, from the member's own copy, once a minute. `server init
+  --patroni` installs it; run it again on a cluster set up before.
+
+The rules files stay, so turning it back on restores them. Nodes added while
+it is off get none. Networks added to `firewall_networks` reach every machine
+the same way. One taken out of it stays let in until it is taken out of
+`/etc/llmgw/firewall.nft` by hand: the rules are only ever added to.
+
+The leading control plane also tells the mirror machine where every node and
+control plane is, so a node on a network the mirror had not let in can fetch
+its weights without a `firewall_networks` entry.
 
 Only those ports are touched. The table's policy is to accept, so whatever
 else the machine serves — and whatever firewall it already has — is left as
@@ -2294,6 +2401,43 @@ A cluster set up online becomes offline the same way: `bundle import` on the
 mirror and control plane, then the `offline:` block in `server.yaml` and a
 restart.
 
+## Stopping and starting the system
+
+From a control plane:
+
+```
+llmgw status                     # does it serve; its control planes, nodes, models, store
+llmgw stop --note "maintenance"  # every model taken down, every request answered 503
+llmgw start                      # every model back as its catalog entry asks
+sudo llmgw restart               # every agent and control plane started again, engines serving throughout
+```
+
+`llmgw stop` frees every GPU and answers requests for a model with 503
+(`system_stopped`, with Retry-After) until `llmgw start`. The control planes,
+agents and mirror keep running, so the console, the admin pages and `llmgw
+start` keep working, and no model's definition changes. It is the cluster's
+setting `serving.stopped`, so a restart does not undo it, and it is in the
+audit with its note. `--wait` waits until every replica is down, or every
+model has its replicas again.
+
+`llmgw restart` restarts every node's agent, then the other control planes
+over ssh (`--ssh-user`, the one sudo was run by), then this one — the leader
+last, so it hands over to a standby already restarted. Engines keep serving
+and are taken over again; nothing is placed anew.
+
+One machine's own services are `llmgw service`:
+
+```
+sudo llmgw service status
+sudo llmgw service stop          # agent, control plane, mirror — whichever this machine has
+sudo llmgw service start
+sudo llmgw service restart agent # or name one: server, agent, mirror
+```
+
+Stopping a node's agent leaves its engines serving — the agent takes them
+over again when it starts — unless `llmgw service stop --engines` stops them
+too, and the control plane then places their replicas elsewhere.
+
 ## Removing llmgw from a machine
 
 ```

docs/knowledge.md

@@ -0,0 +1,408 @@
+# Knowledge bases
+
+A knowledge base is a collection of a company's documents that the cluster
+can answer from. People put their documents into it from their own machines
+with `llmgw-kb`. The tool reads the files on that machine, cuts them into
+passages ("chunks"), has the gateway embed each passage with the person's own
+key, and sends the text and the vectors. The files themselves never leave
+the machine.
+
+## Why retrieval, and not training
+
+Retrieval is the tool for *knowledge*: policies, procedures, contracts. The
+answer can name its source, a changed document is sent again in minutes, and
+a passage only ever reaches people allowed to read its collection.
+
+Training a model on the documents (LoRA) changes how it *writes*, not what it
+reliably *knows*. It has no access control: whatever goes into the weights
+is available to everyone who may use them.
+
+## What it needs
+
+- **The PostgreSQL store with pgvector.** On a SQLite store knowledge bases
+  are off, and `llmgw ctl kb list` and the console say why.
+  - `llmgw server init` and the high-availability database setup install
+    pgvector and enable it.
+  - A cluster set up before this build needs it by hand, once:
+    1. Install the package on every database machine:
+       `sudo apt-get install postgresql-16-pgvector` (use the server's
+       version number).
+    2. On the leader, enable it in llmgw's database:
+       `sudo -u postgres psql -d llmgw -c 'CREATE EXTENSION vector'`.
+    3. Restart the control planes: `llmgw restart`.
+- **An embedding model in the catalog**, with `capabilities.embeddings`.
+  Two good choices handle Turkish and run on CPUs: `BAAI/bge-m3` (1024
+  dimensions) and `Qwen/Qwen3-Embedding-0.6B`.
+
+  Add it as any model is added (**Serve a model** in the console, or the
+  catalog), with its entry saying that it embeds:
+
+  ```yaml
+  capabilities:
+    embeddings: true
+  ```
+
+  A collection stays with the model it was made with: vectors from two
+  models cannot be compared.
+
+  Some embedding models are trained with an instruction before the
+  question and none before the documents. Give it as the entry's
+  `query_prefix`; llmgw puts it before questions only. For
+  Qwen3-Embedding:
+
+  ```yaml
+  query_prefix: "Instruct: Given a web search query, retrieve relevant passages that answer the query\nQuery: "
+  ```
+
+  With pgvector 0.7 or later a collection's index is kept at half
+  precision (`halfvec`), which halves its memory and takes models of up to
+  4000 dimensions; with an older pgvector the limit is 2000. A collection
+  keeps the kind it was made with.
+
+## For administrators
+
+### Make a collection
+
+A collection belongs to one part of the tenancy tree: an organization, a
+team, a project or a person. Whoever is in the owner, at any level below it,
+may read it and add to it.
+
+```sh
+llmgw ctl kb create team:Acme/hr policies bge-m3
+llmgw ctl kb create user:USER-ID notes bge-m3 --org Acme
+```
+
+The model's dimension is asked of the model itself, so the model has to be
+able to serve when the collection is made. `--dims 1024` gives it instead.
+
+In the console: **Knowledge**, then **A**.
+
+### Who may read and add
+
+The owner, and everything below it, may read and add. A **grant** gives the
+same to another part of the same organization, to a person, or to one key:
+
+```sh
+llmgw ctl kb grant Acme/policies project:Acme/sales/crm           # read
+llmgw ctl kb grant Acme/policies team:Acme/legal --write          # read and add
+llmgw ctl kb revoke Acme/policies project:Acme/sales/crm
+```
+
+How access works:
+
+- **Grants add up down the tree.** A grant to a team reaches its projects
+  and their keys.
+- **Nothing crosses organizations.** A grant to another organization's
+  team, project, key or person is refused.
+- **A person's own key** (a key with no project) reads nothing by itself.
+  The person names the project they act in (`llmgw-kb login` asks for it,
+  or `--project`) and must be a member of it. Joining a second team
+  therefore does not quietly widen what every key of theirs can read.
+- **No access looks like no collection.** A collection a key may not read
+  is answered exactly as one that does not exist, so even its name does not
+  leak.
+
+### Answer from collections by default
+
+Clients that cannot add a header still get answers from the documents when
+their part of the tree has a default:
+
+```sh
+llmgw ctl kb default team:Acme/hr Acme/policies Acme/forms
+llmgw ctl kb default key:KEY-ID Acme/policies
+llmgw ctl kb default team:Acme/hr --clear
+```
+
+- **Which default applies.** The narrowest one: a key's over its project's,
+  a project's over its team's, a team's over its organization's.
+- **A request can override it.** A request that names collections uses
+  those instead, and `X-Llmgw-Knowledge: none` turns the default off for
+  that request.
+- **A default never refuses a request.** A collection the caller may not
+  read is left out for them. If the search fails (the embedding model is
+  down, say), the request is answered without passages and the failure is
+  logged.
+- **Changes take effect within 30 seconds.**
+
+In the console: **Knowledge**, Enter on a collection, then **Answer from it
+by default…**.
+
+### Look after it
+
+```sh
+llmgw ctl kb list                          # collections, owners, documents, grants
+llmgw ctl kb docs Acme/policies            # its documents, who added each and when
+llmgw ctl kb rm Acme/policies old/form.pdf # one document out
+llmgw ctl kb search Acme/policies how many days of annual leave
+llmgw ctl kb delete Acme/policies          # the collection, its documents and grants
+```
+
+Rules that apply:
+
+- A project, team or organization that owns a collection is not deleted
+  until its collections are. This is the same rule as for one with members.
+- A person's own collections go with their account (`users delete`,
+  `users forget`).
+- Every document added or removed is in the audit log (`llmgw ctl audit`),
+  by name. The text is never recorded.
+
+## For everyone: `llmgw-kb`
+
+### Get it
+
+Download `llmgw-kb` for your system from the release page, next to `llmgw`:
+
+| System | File |
+|---|---|
+| Windows (Intel/AMD) | `llmgw-kb-VERSION-windows-amd64.exe` |
+| Windows (ARM) | `llmgw-kb-VERSION-windows-arm64.exe` |
+| macOS (Apple silicon) | `llmgw-kb-VERSION-darwin-arm64` |
+| macOS (Intel) | `llmgw-kb-VERSION-darwin-amd64` |
+| Linux | `llmgw-kb-VERSION-linux-amd64` / `-arm64` |
+
+Rename it to `llmgw-kb` (`llmgw-kb.exe` on Windows) and put it somewhere on
+your PATH. You do not need to install anything else.
+
+On macOS, a downloaded program that is not signed is stopped the first time.
+Either right-click it and choose Open, or run
+`xattr -d com.apple.quarantine llmgw-kb`.
+
+### Log in once
+
+```sh
+llmgw-kb login
+```
+
+It asks for three things, checks them, and saves them where only you can
+read them:
+
+- **The gateway's address.**
+- **Your API key.** It is typed without being shown.
+- **A project.** Needed only for a personal key.
+
+Instead of logging in you can set `LLMGW_URL`, `LLMGW_API_KEY` and
+`LLMGW_PROJECT`.
+
+```sh
+llmgw-kb collections        # what your key may read, and which you may add to
+```
+
+### Send your documents
+
+```sh
+llmgw-kb sync policies ~/Documents/HR
+llmgw-kb sync policies "C:\Users\ayse\Documents\İK" --include "*.pdf,*.docx"
+llmgw-kb sync policies ./handbook ./forms --exclude "drafts/**"
+```
+
+Run it as often as you like. The first run sends everything. Later runs
+send only what changed, and do not even read files whose size and time have
+not changed.
+
+| Option | What it does |
+|---|---|
+| `--dry-run` | Reads and cuts everything, sends nothing, and lists what would leave the machine and how much. |
+| `--delete` | Also removes documents whose files are gone. Only those this machine sent, from these folders; never anything someone else sent. |
+| `--retry` | Reads again files skipped before, for example after installing LibreOffice. |
+| `--include`, `--exclude` | Glob patterns. `*.pdf` matches a file name; `reports/**/*.xlsx` matches a path. |
+| `--hidden` | Reads hidden files and folders too (left out by default). |
+| `-v` | Lists unchanged files too. |
+
+Documents are named by their path under the folder, with forward slashes,
+on every system. With several folders, each folder's own name comes first.
+
+```sh
+llmgw-kb status policies    # what this machine sent and skipped, and why
+llmgw-kb history            # the latest runs
+llmgw-kb formats            # what can be read here
+```
+
+### What it reads
+
+| Read here | Formats |
+|---|---|
+| Documents | PDF (text), Word (`.docx`), Excel (`.xlsx`), PowerPoint (`.pptx`, with speaker notes), OpenDocument (`.odt`, `.ods`, `.odp`), RTF, EPUB |
+| Web and mail | HTML, MHTML, e-mail (`.eml`, with its attachments) |
+| Text | Markdown, plain text, CSV, JSON, YAML, XML, subtitles, notebooks (`.ipynb`), source code |
+| Archives | ZIP, with every file inside it read in turn |
+
+| Read with a helper on your machine | Helper |
+|---|---|
+| Word, Excel and PowerPoint 97–2003 (`.doc`, `.xls`, `.ppt`), Pages, Numbers, Keynote and others | [LibreOffice](https://www.libreoffice.org) |
+| Images and scanned PDFs | [Tesseract](https://tesseract-ocr.github.io) (with the Turkish language data), plus Poppler's `pdftoppm` for PDFs |
+| PDFs the built-in reader cannot parse | Poppler's `pdftotext` |
+
+`llmgw-kb formats` says which helpers it found. Text in UTF-8, UTF-16, or
+the Windows Turkish and Western code pages is read correctly.
+
+A file that cannot be read is skipped with the reason, and the run goes on.
+Reasons include a locked Office file, a format with no helper, and a
+scanned page with no OCR.
+
+### What leaves your machine
+
+For each new or changed document, two things leave the machine, both over
+the gateway's TLS connection:
+
+- **The passages, to be embedded.** They go through `/v1/embeddings` like
+  any request of yours. Your key's limits, budget and guardrails apply.
+- **The document itself.** Its name, its file's SHA-256 digest and size,
+  and the passages with their vectors.
+
+The file itself is never sent. `--dry-run` shows exactly what would be.
+
+### The ledger
+
+`llmgw-kb` keeps a small SQLite database of what it did with every
+document: the file it came from, its size, time and digest, and whether it
+was sent, skipped or failed, with the reason.
+
+| System | Where it is |
+|---|---|
+| Windows | `%AppData%\llmgw-kb\ledger.db` |
+| macOS | `~/Library/Application Support/llmgw-kb/ledger.db` |
+| Linux | `~/.config/llmgw-kb/ledger.db` |
+
+Deleting the ledger is safe: the next run hashes every file once, finds what
+the collection already has, and sends nothing that is unchanged. `--ledger`
+uses another file.
+
+## Asking questions of a collection
+
+A chat request names the collections to answer from in a header. Any client
+that can add a header can do it: the OpenAI and Anthropic SDKs, Continue, or
+curl.
+
+```sh
+curl https://llm.example.com/v1/chat/completions \
+  -H "Authorization: Bearer $LLMGW_API_KEY" \
+  -H "X-Llmgw-Knowledge: politikalar" \
+  -d '{"model": "qwen2.5-coder-14b-instruct",
+       "messages": [{"role": "user", "content": "Kaç gün yıllık iznim var?"}]}'
+```
+
+```python
+client.chat.completions.create(
+    model="qwen2.5-coder-14b-instruct",
+    messages=[{"role": "user", "content": "Kaç gün yıllık iznim var?"}],
+    extra_headers={"X-Llmgw-Knowledge": "politikalar"},
+)
+```
+
+What the gateway does with such a request:
+
+1. **It finds the collections.** It looks up the ones named, by name or ID,
+   among those the key may read. A collection it may not read is answered
+   404, the same as one that does not exist.
+2. **It embeds the question.** The last user message is embedded by the
+   collection's model, on the caller's own account: their limits, budget and
+   guardrails apply.
+3. **It searches.** The collections are searched by meaning and by words
+   (below).
+4. **It adds the passages.** The best passages are numbered and added to the
+   system message, with an instruction to answer from them, cite them as
+   [1], [2], and say so when they do not hold the answer. The system message
+   is created if the request has none. Passages that would not fit the
+   model's context are left out, the weakest first.
+5. **It judges and sends the request.** The request, passages included, then
+   goes through the guardrails and to the model as usual. A document
+   carrying instructions aimed at its reader is judged like anything else
+   the model reads.
+
+On the portal's chat page, **Documents** picks a collection the person may
+read in the chosen project, the default, or none. Under each answer, the
+sources it was given are listed by number.
+
+How it is shaped:
+
+| | |
+|---|---|
+| `X-Llmgw-Knowledge` | Collections, comma separated; `none` for no passages, whatever the default. |
+| `X-Llmgw-Knowledge-K` | How many passages, 1 to 20; 5 by default. |
+| `X-Llmgw-Project` | For a personal key: the project it acts in. |
+| `x-llmgw-sources` (answer) | The passages given, as `document#chunk`, in their numbered order; empty when nothing matched. |
+
+Behaviour to know about:
+
+- **No passages matched.** The model is told that nothing in the documents
+  answers the question, and asked to say so rather than answer from general
+  knowledge.
+- **Other APIs.** Requests through `/v1/responses` and `/v1/messages`
+  (Codex, Claude Code) work the same way, since they are answered as chat
+  completions.
+- **Separate prefix caches.** On vLLM replicas, a request given passages
+  carries a `cache_salt` derived from the collections they came from.
+  Requests that could be given the same passages still share cached blocks,
+  and no one else's do: not another organization's, nor another team's of
+  the same one. How fast an answer starts therefore says nothing about what
+  someone else asked.
+- **No answer cache.** Answers from knowledge bases are never served from
+  the answer cache, because two keys of one project can read different
+  collections.
+
+### How a passage is chosen
+
+Each passage scores the following, and the highest scores win:
+
+- its similarity in meaning to the question (cosine, 0 to 1), plus
+- up to 0.3 for the share of the question's words it contains.
+
+Words almost every passage has ("ne", "için", "the", "of") are not counted.
+The meaning does most of the work: "how much leave do fathers get" finds the
+paternity page, not the annual-leave page that shares more of its words.
+The word match helps with what meaning cannot carry, such as a form number,
+a product code or a name.
+
+Metrics: `llmgw_kb_retrievals_total{result}` (found, nothing, refused) and
+`llmgw_kb_retrieval_seconds`.
+
+## Training an adapter on a collection
+
+An adapter (LoRA) changes how a model *writes*: the organisation's words,
+format and tone. It does not reliably teach facts. On the PoC, a small
+model trained on nine policy documents answered 4 of 9 rephrased questions
+right and mixed facts up on the rest, while the same questions through the
+knowledge base got 9 of 9. Use both: the knowledge base for what answers
+say, an adapter for how they sound.
+
+```sh
+# 1. Questions and answers from the collection, written by a model of the
+#    cluster. Read the file through and correct it.
+llmgw ctl kb dataset Acme/policies --model qwen2.5-coder-14b-instruct --out policies.jsonl
+
+# 2. Train on a node. --gpu-memory is what the job needs of a card; it
+#    runs beside an engine when the engine leaves that much free.
+llmgw ctl train start policies-tone --model qwen2.5-0.5b-instruct \
+  --base-repo Qwen/Qwen2.5-0.5B-Instruct --node cuda01 --gpu-memory 2300 --data policies.jsonl
+llmgw ctl train show JOB --follow
+
+# 3. Make it the model's adapter, kept on the mirror.
+llmgw ctl train publish JOB --restart
+```
+
+The node trains with its installed vLLM's torch and a small overlay (peft
+and accelerate), as the engines' user. For a model llama.cpp serves, it also
+converts the adapter to GGUF with the installed llama.cpp's own code. What
+it makes is kept by the control plane (`llmgw ctl train download`).
+
+An adapter is private like its collection:
+
+- **It has an owner.** An adapter trained on a collection carries some of
+  what the collection says in its weights, so it belongs to a part of the
+  tree.
+  - By default the owner is the collection's owner.
+  - `--owner` names another, which must be able to read every collection
+    the pairs came from.
+  - Only callers in the owner's part of the tree see the adapter. To anyone
+    else it is a model that does not exist, absent from `/v1/models` too.
+- **It is public only by choice.** `--public` makes an adapter every
+  caller's. It is refused for pairs from a knowledge base, so it is only for
+  pairs written by hand.
+
+## How it is kept apart
+
+Passages live in PostgreSQL, partitioned by organization, with an index per
+collection. Every search and every write names the organization and the
+collections the caller may use. On top of that, row-level security on the
+table checks the same thing again, so a query that forgot its filter
+returns nothing rather than another tenant's passages.

docs/onboarding.md

@@ -379,16 +379,38 @@ is what someone gets from a borrowed laptop, and the password is what they do no
 have — and it signs every other browser out, which is the point if they are
 changing it because something went wrong.
 
-There is no self-service reset, because this gateway sends no email and a reset
-link with nowhere to send it is not a reset. When someone loses theirs:
+Someone who has forgotten theirs follows **Forgotten your password?** on the
+sign-in page. With mail set up in the cluster's settings, a link to set a new one
+is mailed to their address — the page says the same whether or not the address
+has an account, so nobody can find out who does from it:
+
+```
+llmgw ctl config set mail '{smtp: smtp.example.com:587, from: llmgw@example.com,
+  username: llmgw, password_file: /etc/llmgw/secrets/smtp, base_url: https://llm.example.com}'
+```
+
+`base_url` is where the link points, and it is never taken from the request:
+otherwise a stranger could ask for someone's link from a page of their own. A
+link works once, for an hour, and a newer one replaces it.
+
+Without mail, the page says to ask an operator, who makes the link and hands it
+over however suits:
+
+```
+llmgw users reset-link ayse@example.com --store …    # --base-url without mail.base_url
+```
+
+Or sets a temporary password to give them, and has them change it from their
+profile:
 
 ```
 llmgw users reset-password usr-k4mq… --password "a temporary passphrase" --store …
 ```
 
-Tell them out of band and have them change it from their profile. Their sessions
-are closed — if they lost the password, the reason might be that someone else has
-it. Their API keys keep working; those are separate credentials.
+Either way their sessions are closed — if they lost the password, the reason
+might be that someone else has it. Their API keys keep working; those are
+separate credentials. People who sign in through single sign-on have no password
+here, and get no link.
 
 ## Trying a model in the browser
 

docs/quick-install.md

@@ -52,7 +52,7 @@ install:
 - your administrator account. **The password is shown once. Copy it now.**
 - a model mirror, which downloads model files once for the whole cluster;
 - this machine as a first node, and a small model (Qwen2.5 0.5B) on it;
-- guardrails on every request, with a judge model (Llama Guard 3 1B, two
+- guardrails on every request, with a judge model (Qwen3Guard-Gen 0.6B, two
   replicas) on CPUs — a request it cannot check is refused, not let through — and API keys that expire within 90 days. What is hardened, and how
   to loosen it: [install.md](install.md#2-the-control-plane).
 

docs/reference.md

@@ -18,6 +18,7 @@ rather than repeated.
 | `name` | `string` | [Adapter](#adapter) |
 | `source` | `string` | [Adapter](#adapter) |
 | `scale` | `float64` | [Adapter](#adapter) |
+| `owner` | `string` | [Adapter](#adapter) |
 | `enabled` | `bool` | [Admission](#admission) |
 | `bands` | `[]Band` | [Admission](#admission) |
 | `default_band` | `string` | [Admission](#admission) |
@@ -26,6 +27,8 @@ rather than repeated.
 | `queue_capacity_per_band` | `int` | [Admission](#admission) |
 | `queue_capacity_total` | `int` | [Admission](#admission) |
 | `interval` | `time.Duration` | [Admission](#admission) |
+| `default` | `map[string]any` | [AgentSettings](#agentsettings) |
+| `nodes` | `map[string]map[string]any` | [AgentSettings](#agentsettings) |
 | `enabled` | `bool` | [Auth](#auth) |
 | `store` | `string` | [Auth](#auth) |
 | `cache_ttl` | `time.Duration` | [Auth](#auth) |
@@ -57,6 +60,7 @@ rather than repeated.
 | `tools` | `bool` | [Capabilities](#capabilities) |
 | `vision` | `bool` | [Capabilities](#capabilities) |
 | `embeddings` | `bool` | [Capabilities](#capabilities) |
+| `images` | `bool` | [Capabilities](#capabilities) |
 | `interval` | `time.Duration` | [Control](#control) |
 | `base_backoff` | `time.Duration` | [Control](#control) |
 | `max_backoff` | `time.Duration` | [Control](#control) |
@@ -73,6 +77,7 @@ rather than repeated.
 | `api_key` | `string` | [Endpoint](#endpoint) |
 | `models` | `[]string` | [Endpoint](#endpoint) |
 | `engine` | `string` | [Endpoint](#endpoint) |
+| `workflow` | `string` | [Endpoint](#endpoint) |
 | `pool` | `string` | [Endpoint](#endpoint) |
 | `weight` | `float64` | [Endpoint](#endpoint) |
 | `priority` | `int` | [Endpoint](#endpoint) |
@@ -87,6 +92,8 @@ rather than repeated.
 | `model` | `string` | [External](#external) |
 | `api_key_file` | `string` | [External](#external) |
 | `peer` | `string` | [External](#external) |
+| `api` | `string` | [External](#external) |
+| `workflow` | `string` | [External](#external) |
 | `budgets` | `[]FacilityBudget` | [Facility](#facility) |
 | `margin` | `float64` | [Facility](#facility) |
 | `gpu_watts` | `float64` | [Facility](#facility) |
@@ -102,9 +109,11 @@ rather than repeated.
 | `format` | `string` | [GuardModel](#guardmodel) |
 | `categories` | `[]string` | [GuardModel](#guardmodel) |
 | `model` | `string` | [JudgePolicy](#judgepolicy) |
+| `models` | `[]string` | [JudgePolicy](#judgepolicy) |
 | `input` | `string` | [JudgePolicy](#judgepolicy) |
 | `output` | `string` | [JudgePolicy](#judgepolicy) |
 | `categories` | `[]string` | [JudgePolicy](#judgepolicy) |
+| `controversial` | `string` | [JudgePolicy](#judgepolicy) |
 | `latency_budget` | `time.Duration` | [JudgePolicy](#judgepolicy) |
 | `tokenize_timeout` | `time.Duration` | [KVIndex](#kvindex) |
 | `min_blocks` | `int` | [KVIndex](#kvindex) |
@@ -140,6 +149,7 @@ rather than repeated.
 | `context_length` | `int` | [Model](#model) |
 | `max_output` | `int` | [Model](#model) |
 | `capabilities` | `Capabilities` | [Model](#model) |
+| `query_prefix` | `string` | [Model](#model) |
 | `role` | `string` | [Model](#model) |
 | `guard` | `*GuardModel` | [Model](#model) |
 | `origin` | `*Origin` | [Model](#model) |
@@ -203,6 +213,7 @@ rather than repeated.
 | `files` | `[]OriginFile` | [PreviousOrigin](#previousorigin) |
 | `name` | `string` | [Profile](#profile) |
 | `engine` | `string` | [Profile](#profile) |
+| `engine_version` | `string` | [Profile](#profile) |
 | `gpu_class` | `string` | [Profile](#profile) |
 | `tp` | `int` | [Profile](#profile) |
 | `pp` | `int` | [Profile](#profile) |
@@ -288,7 +299,10 @@ rather than repeated.
 | `guard` | `guard.Config` | [Server](#server) |
 | `secrets` | `Secrets` | [Server](#server) |
 | `notify` | `notify.Config` | [Server](#server) |
+| `mail` | `mail.Config` | [Server](#server) |
+| `serving` | `ServingSettings` | [Server](#server) |
 | `firewall` | `FirewallSettings` | [Server](#server) |
+| `agents` | `AgentSettings` | [Server](#server) |
 | `firewall_networks` | `[]string` | [Server](#server) |
 | `backup_recipients` | `[]string` | [Server](#server) |
 | `backup_copy_to` | `[]string` | [Server](#server) |
@@ -300,6 +314,7 @@ rather than repeated.
 | `facility` | `Facility` | [Server](#server) |
 | `federation` | `Federation` | [Server](#server) |
 | `endpoints` | `[]Endpoint` | [Server](#server) |
+| `stopped` | `bool` | [ServingSettings](#servingsettings) |
 | `bin` | `string` | [SimConfig](#simconfig) |
 | `python` | `string` | [TRTLLMConfig](#trtllmconfig) |
 | `store` | `string` | [Threads](#threads) |
@@ -372,8 +387,14 @@ rather than repeated.
 | `llamacpp` | `LlamaCppConfig` | [engines.Config](#enginesconfig) |
 | `trtllm` | `TRTLLMConfig` | [engines.Config](#enginesconfig) |
 | `sim` | `SimConfig` | [engines.Config](#enginesconfig) |
+| `enabled` | `*bool` | [guard.Config](#guardconfig) |
 | `default` | `[]string` | [guard.Config](#guardconfig) |
 | `policies` | `map[string]Policy` | [guard.Config](#guardconfig) |
+| `smtp` | `string` | [mail.Config](#mailconfig) |
+| `from` | `string` | [mail.Config](#mailconfig) |
+| `username` | `string` | [mail.Config](#mailconfig) |
+| `password_file` | `string` | [mail.Config](#mailconfig) |
+| `base_url` | `string` | [mail.Config](#mailconfig) |
 | `webhooks` | `[]Webhook` | [notify.Config](#notifyconfig) |
 | `host` | `string` | [notify.Config](#notifyconfig) |
 | `work_dir` | `string` | [supervisor.Config](#supervisorconfig) |
@@ -411,6 +432,16 @@ weights are.
 
 Scale is how strongly it is applied; zero takes 1.
 
+### `owner`
+
+`string`
+
+Owner is the part of the tenancy tree the adapter belongs to —
+org:ID, team:ID, project:ID, user:ID or key:ID — and only callers in
+it see it: to anyone else it is a model that does not exist. An
+adapter trained on an organisation's documents carries some of what
+they say in its weights. Empty is every caller the model allows.
+
 ## Admission
 
 ### `enabled`
@@ -457,6 +488,21 @@ Zero takes half of all the bands' capacities together.
 
 `time.Duration`
 
+## AgentSettings
+
+### `default`
+
+`map[string]any`
+
+Default is agent.yaml's keys for every node.
+
+### `nodes`
+
+`map[string]map[string]any`
+
+Nodes are agent.yaml's keys for one node each, by node id, over
+Default.
+
 ## Auth
 
 ### `enabled`
@@ -674,6 +720,14 @@ token is MaxTTFTRatio times this model's. Zero is 50, 0.05 and 1.5.
 
 `bool`
 
+### `images`
+
+`bool`
+
+Images is a model that makes pictures, asked for at
+/v1/images/generations. Implied by an external model whose API is
+ComfyUI's.
+
 ## Control
 
 ### `interval`
@@ -787,6 +841,13 @@ with one.
 Engine names the inference engine, which decides what its metrics are
 called. Empty means vLLM's names.
 
+### `workflow`
+
+`string`
+
+Workflow is the ComfyUI workflow for engine: comfyui, a built-in's
+name or its JSON; flux-schnell unless set.
+
 ### `pool`
 
 `string`
@@ -878,6 +939,23 @@ a model that federation added (`llmgw ctl federation add`) rather
 than one written by hand. Such a model is named PEER/MODEL, comes and
 goes with the peer's own list, and is sent only while the peer is up.
 
+### `api`
+
+`string`
+
+API is what the provider speaks: OpenAI's, unless this says comfyui —
+a ComfyUI server, asked for pictures at /v1/images/generations and
+sent the workflow below with the request's prompt in it. Model is then
+the checkpoint file the workflow loads, as ComfyUI lists it.
+
+### `workflow`
+
+`string`
+
+Workflow is the ComfyUI workflow, in its API format: a built-in's name
+(flux-schnell, the default, or flux-dev) or a JSON file on the control
+plane.
+
 ## Facility
 
 ### `budgets`
@@ -984,9 +1062,11 @@ every node and control plane within seconds.
 
 `string`
 
-Format is llamaguard3 (Meta's Llama Guard 3: "safe", or "unsafe" and
-the categories), shieldgemma (Google's ShieldGemma: Yes or No), or
-json ({"safe": false, "categories": ["..."]}).
+Format is qwen3guard (Qwen3Guard-Gen: "Safety: Safe", "Unsafe" or
+"Controversial", and the categories), llamaguard3 (Meta's Llama Guard
+3: "safe", or "unsafe" and the categories), shieldgemma (Google's
+ShieldGemma: Yes or No), or json ({"safe": false, "categories":
+["..."]}).
 
 ### `categories`
 
@@ -1003,6 +1083,15 @@ sentence — and the names reported for llamaguard3's S1, S2, ….
 
 Model is a catalog model with role: guard.
 
+### `models`
+
+`[]string`
+
+Models are judges asked in turn, instead of Model: the first that
+answers gives the verdict. A judge on a GPU first and one on CPUs
+after it is the GPU's speed while it is there, and the CPUs' when it
+is not — busy, failed, or the node down.
+
 ### `input`
 
 `string`
@@ -1021,12 +1110,19 @@ Empty does not ask the judge about that direction.
 Categories, when set, are the only ones that count: Llama Guard's
 codes (S1, S10) or the names the model reports.
 
+### `controversial`
+
+`string`
+
+Controversial is what Qwen3Guard's third verdict counts as: safe
+(the default) or unsafe.
+
 ### `latency_budget`
 
 `time.Duration`
 
-LatencyBudget is how long the judge is waited for before the policy's
-fail applies. Zero is 250ms.
+LatencyBudget is how long each judge is waited for before the next is
+asked, or, after the last, the policy's fail applies. Zero is 250ms.
 
 ## KVIndex
 
@@ -1330,6 +1426,16 @@ Capabilities is what this model can do beyond plain completion.
 
 Takes a [Capabilities](#capabilities) block.
 
+### `query_prefix`
+
+`string`
+
+QueryPrefix is put before a question an embedding model embeds for a
+knowledge-base search; documents are embedded without it. Models
+trained with an instruction on the query side want it, Qwen3-Embedding
+"Instruct: Given a web search query, retrieve relevant passages that
+answer the query\nQuery: " for one.
+
 ### `role`
 
 `string`
@@ -1755,6 +1861,14 @@ before affinity is abandoned.
 
 `string`
 
+### `engine_version`
+
+`string`
+
+EngineVersion restricts the profile to nodes whose installed engine's
+version starts with it: "0.11" takes vLLM 0.11.0 and 0.11.1, not
+0.10.2. Empty takes whatever version the node has.
+
 ### `gpu_class`
 
 `string`
@@ -2456,6 +2570,24 @@ Prometheus and Alertmanager in between. Empty sends nothing.
 
 Takes a [notify.Config](#notifyconfig) block.
 
+### `mail`
+
+`mail.Config`
+
+Mail is where the gateway's few emails go out from: a link to set a
+forgotten password. Unset, none is sent and people ask an operator.
+
+Takes a [mail.Config](#mailconfig) block.
+
+### `serving`
+
+`ServingSettings`
+
+Serving stops and starts the whole cluster's service (llmgw stop, llmgw
+start).
+
+Takes a [ServingSettings](#servingsettings) block.
+
 ### `firewall`
 
 `FirewallSettings`
@@ -2464,6 +2596,18 @@ Firewall turns llmgw's firewall on every machine on or off.
 
 Takes a [FirewallSettings](#firewallsettings) block.
 
+### `agents`
+
+`AgentSettings`
+
+Agents are the nodes' agent settings: agent.yaml's keys but those an
+agent needs to start, for every node (default) and for one
+(nodes.<id>). Sent to the agents as they change; an agent whose
+settings changed starts again on them, its engines serving
+throughout.
+
+Takes a [AgentSettings](#agentsettings) block.
+
 ### `firewall_networks`
 
 `[]string`
@@ -2554,6 +2698,18 @@ Endpoints is the phase-1 static replica list. It is only consulted when
 no model declares a profile, which is what makes a router-only smoke test
 against sim-engines possible without a control plane.
 
+## ServingSettings
+
+### `stopped`
+
+`bool`
+
+Stopped takes every model down and answers every request for one with
+503 (system_stopped), leaving the control planes, agents and mirror
+running and every model's definition as it is: llmgw stop sets it,
+llmgw start clears it, and the models come back as their catalog
+entries ask. In force on every control plane within seconds.
+
 ## SimConfig
 
 ### `bin`
@@ -3067,6 +3223,13 @@ Takes a [SimConfig](#simconfig) block.
 
 ## guard.Config
 
+### `enabled`
+
+`*bool`
+
+Enabled turns every policy on or off at once: false checks nothing,
+whatever is defined or attached. Unset is on.
+
 ### `default`
 
 `[]string`
@@ -3075,6 +3238,40 @@ Takes a [SimConfig](#simconfig) block.
 
 `map[string]Policy`
 
+## mail.Config
+
+### `smtp`
+
+`string`
+
+SMTP is the relay, host:port: 587 with STARTTLS, 465 with TLS from the
+start, 25 as the relay allows. Empty sends nothing.
+
+### `from`
+
+`string`
+
+From is the sender, llmgw@example.com or "llmgw <llmgw@example.com>".
+
+### `username`
+
+`string`
+
+Username and PasswordFile log in to the relay, when it wants that.
+The password is a secret reference: a file, env:NAME or vault:path.
+
+### `password_file`
+
+`string`
+
+### `base_url`
+
+`string`
+
+BaseURL is the gateway's public address, which links in mail point
+at: https://llm.example.com. Never taken from a request, whose Host a
+sender chooses and could point a reset link at their own site.
+
 ## notify.Config
 
 ### `webhooks`

go.mod

@@ -18,6 +18,7 @@ require (
 	github.com/knadh/koanf/providers/env/v2 v2.0.1
 	github.com/knadh/koanf/providers/file v1.2.1
 	github.com/knadh/koanf/v2 v2.3.6
+	github.com/ledongthuc/pdf v0.0.0-20260907135840-6c8c28e0e8a0
 	github.com/mattn/go-isatty v0.0.24
 	github.com/prometheus/client_golang v1.24.1
 	github.com/prometheus/client_model v0.6.3
@@ -34,6 +35,7 @@ require (
 	golang.org/x/net v0.59.0
 	golang.org/x/oauth2 v0.37.0
 	golang.org/x/sync v0.23.0
+	golang.org/x/term v0.46.0
 	golang.org/x/text v0.42.0
 	google.golang.org/protobuf v1.36.12
 	modernc.org/sqlite v1.59.0

go.sum

@@ -95,6 +95,8 @@ github.com/knadh/koanf/v2 v2.3.6 h1:JoQPSJmvS4aP0xNc8xMDr5tcrkSEInL23/Il7pITAKo=
 github.com/knadh/koanf/v2 v2.3.6/go.mod h1:gRb40VRAbd4iJMYYD5IxZ6hfuopFcXBpc9bbQpZwo28=
 github.com/kylelemons/godebug v1.1.0 h1:RPNrshWIDI6G2gRW9EHilWtl7Z6Sb1BR0xunSBf0SNc=
 github.com/kylelemons/godebug v1.1.0/go.mod h1:9/0rRGxNHcop5bhtWyNeEfOS8JIWk580+fNqagV/RAw=
+github.com/ledongthuc/pdf v0.0.0-20260907135840-6c8c28e0e8a0 h1:7Q+xNAZFmnfYOMweHN3c/PDFUKKfY1pVJ26K++QvVfU=
+github.com/ledongthuc/pdf v0.0.0-20260907135840-6c8c28e0e8a0/go.mod h1:1fEHWurg7pvf5SG6XNE5Q8UZmOwex51Mkx3SLhrW5B4=
 github.com/lucasb-eyer/go-colorful v1.3.0 h1:2/yBRLdWBZKrf7gB40FoiKfAWYQ0lqNcbuQwVHXptag=
 github.com/lucasb-eyer/go-colorful v1.3.0/go.mod h1:R4dSotOR9KMtayYi1e77YzuveK+i7ruzyGqttikkLy0=
 github.com/mattn/go-isatty v0.0.24 h1:tGZZoVgT/KiqK1c8ocVLeDS8BSWMRd47J3Lbz7vsReI=

internal/account/account_test.go

@@ -448,3 +448,40 @@ func TestAHashThatCannotGetATurnSaysTheServerIsBusy(t *testing.T) {
 		t.Errorf("a hash with no turn: %v", err)
 	}
 }
+
+func TestAForgottenPasswordIsResetOnceThroughALink(t *testing.T) {
+	svc, st := newService(t)
+	ctx := context.Background()
+	must(t, svc.SignUp(ctx, "ayse@example.com", "A", goodPassword))
+	u, _ := st.UserByEmail(ctx, "ayse@example.com")
+
+	// Waiting for approval: no link, said the same as for nobody.
+	if _, _, err := svc.RequestReset(ctx, st, "ayse@example.com"); !errors.Is(err, ErrNoReset) {
+		t.Fatalf("a pending account got a link: %v", err)
+	}
+	if _, _, err := svc.RequestReset(ctx, st, "nobody@example.com"); !errors.Is(err, ErrNoReset) {
+		t.Fatalf("an unknown address: %v", err)
+	}
+	must(t, st.ApproveUser(ctx, u.ID, "proj", "ilker", time.Now()))
+	session, _, err := svc.SignIn(ctx, "ayse@example.com", goodPassword)
+	must(t, err)
+
+	token, got, err := svc.RequestReset(ctx, st, "AYSE@example.com")
+	if err != nil || token == "" || got.ID != u.ID {
+		t.Fatalf("an active account: %q %+v %v", token, got, err)
+	}
+	if err := svc.ResetPassword(ctx, st, st, token, "short"); !errors.Is(err, ErrWeakPassword) {
+		t.Fatalf("a weak password: %v", err)
+	}
+	next := goodPassword + "-new"
+	must(t, svc.ResetPassword(ctx, st, st, token, next))
+	if _, _, err := svc.Resolve(ctx, session); err == nil {
+		t.Error("a session from before the reset still works")
+	}
+	if _, _, err := svc.SignIn(ctx, "ayse@example.com", next); err != nil {
+		t.Errorf("the new password does not sign in: %v", err)
+	}
+	if err := svc.ResetPassword(ctx, st, st, token, next+"2"); !errors.Is(err, ErrNoReset) {
+		t.Errorf("a link used twice: %v", err)
+	}
+}

internal/account/service.go

@@ -373,3 +373,63 @@ func (s *Service) OpenSession(ctx context.Context, userID string) (token, csrf s
 	}
 	return token, csrf, nil
 }
+
+// ResetLinkTTL is how long a password reset link works.
+const ResetLinkTTL = time.Hour
+
+// ResetStore keeps password reset links.
+type ResetStore interface {
+	UserByEmail(ctx context.Context, email string) (store.User, error)
+	CreatePasswordReset(ctx context.Context, userID string, ttl time.Duration) (string, error)
+	PasswordResetUser(ctx context.Context, token string) (string, error)
+	UsePasswordReset(ctx context.Context, token string) (string, error)
+}
+
+// ErrNoReset is an address that gets no reset link: none has an account, the
+// account signs in through single sign-on, or it is not active. Never told to
+// whoever asked, who learns the same either way.
+var ErrNoReset = errors.New("no reset link for that address")
+
+// RequestReset mints a reset link token for the person an address belongs
+// to, when they sign in with a password and are active. A few a quarter of
+// an hour per address: each sends a mail.
+func (s *Service) RequestReset(ctx context.Context, rs ResetStore, email string) (string, store.User, error) {
+	email = store.NormalizeEmail(email)
+	if !validEmail(email) {
+		return "", store.User{}, ErrNoReset
+	}
+	if !s.allowAttempt("reset:"+email, 3) {
+		return "", store.User{}, ErrTooManyTries
+	}
+	u, err := rs.UserByEmail(ctx, email)
+	if err != nil || u.PasswordHash == "" || u.State != store.UserActive {
+		return "", store.User{}, ErrNoReset
+	}
+	token, err := rs.CreatePasswordReset(ctx, u.ID, ResetLinkTTL)
+	if err != nil {
+		return "", store.User{}, err
+	}
+	s.log.Info("a password reset link was made", "user", u.ID)
+	return token, u, nil
+}
+
+// ResetPassword sets a password from a reset link, which is spent, and signs
+// the person out everywhere: if they lost it, someone else may have it.
+func (s *Service) ResetPassword(ctx context.Context, rs ResetStore, pw PasswordChanger, token, next string) error {
+	hash, err := HashPassword(next)
+	if err != nil {
+		return err
+	}
+	userID, err := rs.UsePasswordReset(ctx, token)
+	if err != nil {
+		return ErrNoReset
+	}
+	if err := pw.SetPassword(ctx, userID, hash); err != nil {
+		return err
+	}
+	if err := pw.DeleteUserSessions(ctx, userID); err != nil {
+		return err
+	}
+	s.log.Info("a password was set from a reset link", "user", userID)
+	return nil
+}

internal/agent/agent.go

@@ -34,6 +34,7 @@ import (
 	"gitea.mynodes.xyz/ilker/llmgw/internal/agent/engines"
 	"gitea.mynodes.xyz/ilker/llmgw/internal/agent/provision"
 	"gitea.mynodes.xyz/ilker/llmgw/internal/agent/supervisor"
+	"gitea.mynodes.xyz/ilker/llmgw/internal/agent/train"
 	"gitea.mynodes.xyz/ilker/llmgw/internal/agent/weights"
 	"gitea.mynodes.xyz/ilker/llmgw/internal/control/agentsvc"
 	"gitea.mynodes.xyz/ilker/llmgw/internal/firewall"
@@ -136,6 +137,13 @@ type Agent struct {
 	// by mu.
 	kv     *kvevents.Mirror
 	kvSubs map[string]*kvevents.Subscriber
+
+	// settingsCheck says whether the cluster's settings for this agent
+	// would load beside its agent.yaml; see SetSettingsCheck.
+	settingsCheck func([]byte) error
+
+	// trainer runs LoRA training jobs, one at a time (BACKLOG F4).
+	trainer *train.Runner
 }
 
 // connectHTTPClient is the HTTP client connect uses; an interface so tests can
@@ -182,6 +190,18 @@ func New(cfg Config, httpClient connectHTTPClient, log *slog.Logger) (*Agent, er
 		kvSubs:       map[string]*kvevents.Subscriber{},
 	}
 	a.prov = &provision.Installer{Root: cfg.EnginesDir, Log: log, OnStatus: a.provisionStatus}
+	runAs := cfg.Superv.RunAs
+	if runAs == "" {
+		runAs = supervisor.DefaultRunAs
+	}
+	// Where the engines are, as the installer has it when the file says
+	// nothing.
+	enginesDir := cfg.EnginesDir
+	if enginesDir == "" {
+		enginesDir = provision.DefaultRoot
+	}
+	a.trainer = train.New(train.Env{EnginesDir: enginesDir, WorkDir: filepath.Join(cfg.Superv.WorkDir, "trainer"),
+		RunAs: runAs, Log: log})
 	return a, nil
 }
 
@@ -371,6 +391,12 @@ func (a *Agent) settleAdopted(sh *supervisor.Shard, o supervisor.Orphan) {
 		_ = a.sup.Stop(ctx, o.ShardID)
 		return
 	}
+	// What it computes on, for one started before the agent kept that.
+	if a.sup.ComputeOf(o.ShardID) == nil {
+		if c, ok := a.gpuCompute(ctx, o.Spec.GPUIndices, sh.Status().PID); ok {
+			a.sup.SetCompute(o.ShardID, c)
+		}
+	}
 	a.log.Info("adopted engine is serving",
 		"shard", o.ShardID, "replica", o.ReplicaID)
 }
@@ -411,6 +437,17 @@ func (a *Agent) weights() *weights.Fetcher {
 }
 
 // runningShards describes what this agent is supervising, for registration.
+// unready are the shards a registration reported as not yet ready.
+func unready(running []*llmgwv1.RunningShard) map[string]bool {
+	out := map[string]bool{}
+	for _, r := range running {
+		if r.GetState() != agentsvc.StateToWire(supervisor.Ready) {
+			out[r.GetShardId()] = true
+		}
+	}
+	return out
+}
+
 func (a *Agent) runningShards() []*llmgwv1.RunningShard {
 	var out []*llmgwv1.RunningShard
 	for _, st := range a.sup.Statuses() {
@@ -436,6 +473,7 @@ func (a *Agent) runningShards() []*llmgwv1.RunningShard {
 			Pool:       spec.Pool,
 			Fabric:     spec.Fabric,
 			ProfileFp:  spec.ProfileFP,
+			Compute:    computeToWire(a.sup.ComputeOf(st.ID)),
 			LeaderAddr: a.cfg.MgmtAddr,
 			LeaderPort: int32(spec.LeaderPort),
 		})
@@ -564,8 +602,9 @@ func (a *Agent) session(ctx context.Context, server string) error {
 		a.sendLoop(ctx, stream, out)
 	}()
 	defer wg.Wait()
-	go a.watchExits(ctx, out)
+	go a.watchExits(ctx, out, unready(reg.RunningShards))
 	go a.forwardKV(ctx, out)
+	go a.reportCompute(ctx, out)
 
 	for {
 		msg, err := stream.Receive()
@@ -861,6 +900,20 @@ func (a *Agent) handle(ctx context.Context, msg *llmgwv1.ControlMessage, out cha
 		a.setControlPlanes(m.ControlPlanes)
 	case *llmgwv1.ControlMessage_Firewall:
 		go a.applyFirewall(m.Firewall)
+	case *llmgwv1.ControlMessage_AgentSettings:
+		go a.applySettings(m.AgentSettings)
+	case *llmgwv1.ControlMessage_TrainStart:
+		go a.trainJob(ctx, m.TrainStart, out)
+	case *llmgwv1.ControlMessage_TrainCancel:
+		a.trainer.Cancel(m.TrainCancel.GetJobId())
+	case *llmgwv1.ControlMessage_RestartAgent:
+		go func() {
+			a.log.Info("restarting, as the control plane asks; the engines keep serving",
+				"reason", m.RestartAgent.GetReason())
+			// A moment for the log line to leave before the process does.
+			time.Sleep(300 * time.Millisecond)
+			exitForUpgrade()
+		}()
 	case *llmgwv1.ControlMessage_Logs:
 		reply := a.logsReply(m.Logs)
 		// Off the receive loop: a full send queue must not stop commands being
@@ -1130,6 +1183,14 @@ func (a *Agent) startShard(ctx context.Context, m *llmgwv1.StartShard, out chan<
 	} else {
 		err = a.sup.WaitAlive(ctx, sh)
 	}
+	// Where it put the model, by its own account: said with the readiness,
+	// and a GPU placement it computes on the CPU refused even where
+	// nvidia-smi cannot be asked.
+	compute, said, cerr := engineCompute(eng, spec.GPUIndices, sh.Logs(512))
+	if err == nil && cerr != nil {
+		err = cerr
+		_ = a.sup.Stop(ctx, shardID)
+	}
 	if err == nil {
 		err = a.onItsGPUs(ctx, eng, spec, sh.Status().PID)
 		if err != nil {
@@ -1142,7 +1203,19 @@ func (a *Agent) startShard(ctx context.Context, m *llmgwv1.StartShard, out chan<
 		report(supervisor.Failed, err.Error(), sh.Logs(20))
 		return
 	}
-	report(supervisor.Ready, "", nil)
+	// The engine's own account when it gave one, with its layers; else what
+	// nvidia-smi sees the shard holding.
+	if !said {
+		compute, said = a.gpuCompute(ctx, spec.GPUIndices, sh.Status().PID)
+	}
+	if said {
+		// Kept with the shard, so an agent restarted, or a control plane
+		// taking over, still learns it.
+		a.sup.SetCompute(shardID, compute)
+		send(&llmgwv1.ShardStatus{State: agentsvc.StateToWire(supervisor.Ready), Compute: computeToWire(&compute)})
+	} else {
+		report(supervisor.Ready, "", nil)
+	}
 	// Its dying later is reported by the session's watchExits.
 }
 
@@ -1155,7 +1228,14 @@ func (a *Agent) startShard(ctx context.Context, m *llmgwv1.StartShard, out chan<
 // reported the engine dying; and an adopted engine had no watcher at all —
 // on node06 one was killed and its replica stayed "ready" for the control
 // plane, routed to and failing, indefinitely.
-func (a *Agent) watchExits(ctx context.Context, out chan<- *llmgwv1.AgentMessage) {
+//
+// It also reports the shards the registration said were not ready yet once
+// they are. An adopted engine is health-checked after the agent registers,
+// and a shard started on the stream before is reported on that stream; so
+// when the registration came first nothing ever told the control plane, and
+// the replica stayed "starting" while it served — on node05, after a
+// control plane restart, the only replica of a model.
+func (a *Agent) watchExits(ctx context.Context, out chan<- *llmgwv1.AgentMessage, pending map[string]bool) {
 	t := time.NewTicker(time.Second)
 	defer t.Stop()
 	ready := map[string]bool{}
@@ -1164,6 +1244,24 @@ func (a *Agent) watchExits(ctx context.Context, out chan<- *llmgwv1.AgentMessage
 			switch st.State {
 			case supervisor.Ready:
 				ready[st.ID] = true
+				if !pending[st.ID] {
+					continue
+				}
+				delete(pending, st.ID)
+				sh, err := a.sup.Get(st.ID)
+				if err != nil {
+					continue
+				}
+				msg := &llmgwv1.AgentMessage{Msg: &llmgwv1.AgentMessage_Shard{Shard: &llmgwv1.ShardStatus{
+					NodeId: a.cfg.NodeID, ReplicaId: sh.Spec.ReplicaID, ShardId: st.ID,
+					State:   agentsvc.StateToWire(st.State),
+					Compute: computeToWire(a.sup.ComputeOf(st.ID)),
+				}}}
+				select {
+				case out <- msg:
+				case <-ctx.Done():
+					return
+				}
 			case supervisor.Failed:
 				if !ready[st.ID] {
 					continue // never ready here: its start reported it
@@ -1373,3 +1471,42 @@ func ResolveManagement(c Config) (addr, iface string) {
 	}
 	return addr, iface
 }
+
+// reportCompute tells the control plane, once a session, what each running
+// shard computes on as soon as that is known: an adopted engine's is found
+// after registration, which carried what was known then.
+func (a *Agent) reportCompute(ctx context.Context, out chan<- *llmgwv1.AgentMessage) {
+	sent := map[string]bool{}
+	t := time.NewTicker(5 * time.Second)
+	defer t.Stop()
+	for range 60 {
+		for _, st := range a.sup.Statuses() {
+			if st.State.Terminal() || sent[st.ID] {
+				continue
+			}
+			c := a.sup.ComputeOf(st.ID)
+			if c == nil {
+				continue
+			}
+			sh, err := a.sup.Get(st.ID)
+			if err != nil {
+				continue
+			}
+			// No state: a report of compute alone, not a transition.
+			msg := &llmgwv1.AgentMessage{Msg: &llmgwv1.AgentMessage_Shard{Shard: &llmgwv1.ShardStatus{
+				NodeId: a.cfg.NodeID, ReplicaId: sh.Spec.ReplicaID, ShardId: st.ID, Compute: computeToWire(c),
+			}}}
+			select {
+			case out <- msg:
+				sent[st.ID] = true
+			case <-ctx.Done():
+				return
+			}
+		}
+		select {
+		case <-ctx.Done():
+			return
+		case <-t.C:
+		}
+	}
+}

internal/agent/agent_test.go

@@ -37,11 +37,12 @@ import (
 type statusRecorder struct {
 	ch chan status
 
-	mu      sync.Mutex
-	calls   []adoptCall
-	adopted []reconciler.AdoptedShard
-	unclaim []string // shard ids to report back as unclaimed
-	fetches []fetchReport
+	mu       sync.Mutex
+	calls    []adoptCall
+	adopted  []reconciler.AdoptedShard
+	unclaim  []string // shard ids to report back as unclaimed
+	fetches  []fetchReport
+	computes []reconciler.Compute
 }
 
 type status struct {
@@ -61,6 +62,12 @@ func (r *statusRecorder) OnShardStatus(replicaID, shardID string, st supervisor.
 
 // fetches records download progress, so a test can assert that a fetch is
 // visible from the control plane rather than only in the node's log.
+func (r *statusRecorder) OnShardCompute(replicaID, shardID string, c reconciler.Compute) {
+	r.mu.Lock()
+	defer r.mu.Unlock()
+	r.computes = append(r.computes, c)
+}
+
 func (r *statusRecorder) OnFetchProgress(replicaID, shardID, source string, done, total int64) {
 	r.mu.Lock()
 	defer r.mu.Unlock()

internal/agent/config.go

@@ -103,6 +103,11 @@ type Config struct {
 	Tracing obs.Tracing `koanf:"tracing"`
 
 	Version string
+
+	// Settings is the cluster's settings for this agent it was started on,
+	// as the control plane sent them: what a newly sent version is
+	// compared with.
+	Settings string `koanf:"-"`
 }
 
 // DefaultConfig returns sensible defaults.

internal/agent/engines/compute.go

@@ -0,0 +1,79 @@
+package engines
+
+import (
+	"regexp"
+	"strconv"
+	"strings"
+)
+
+// Compute is where an engine runs its model, as it said while loading it.
+type Compute struct {
+	// Backend is CUDA, Vulkan, Metal, ROCm or CPU.
+	Backend string
+	// Devices are those it named: "CUDA0: Quadro RTX 6000 (7.5)".
+	Devices []string
+	// Offloaded of Total layers went to the devices; 0 and 0 when it did
+	// not say.
+	Offloaded, Total int
+	// NoGPU is the engine saying it found no device it could use.
+	NoGPU bool
+}
+
+var (
+	offloadedRE = regexp.MustCompile(`offloaded (\d+)/(\d+) layers to GPU`)
+	usingRE     = regexp.MustCompile(`using device ([A-Za-z]+\d+) \(([^)]*)\)`)
+	ccRE        = regexp.MustCompile(`Device (\d+): ([^,]+), compute capability ([0-9.]+)`)
+	noDevicesRE = regexp.MustCompile(`found 0 (CUDA|Vulkan|ROCm) devices|no usable GPU found`)
+)
+
+// ReadLlamaCompute reads llama.cpp's start-up output for where it put the
+// model, reporting false when the output says nothing of it.
+func ReadLlamaCompute(lines []string) (Compute, bool) {
+	var c Compute
+	known := false
+	cc := map[string]string{} // device index → compute capability
+	for _, l := range lines {
+		if m := ccRE.FindStringSubmatch(l); m != nil {
+			cc[m[1]] = m[3]
+		}
+	}
+	seen := map[string]bool{}
+	for _, l := range lines {
+		if m := offloadedRE.FindStringSubmatch(l); m != nil {
+			c.Offloaded, _ = strconv.Atoi(m[1])
+			c.Total, _ = strconv.Atoi(m[2])
+			known = true
+		}
+		if m := usingRE.FindStringSubmatch(l); m != nil && !seen[m[1]] {
+			seen[m[1]] = true
+			d := m[1] + ": " + strings.TrimSpace(m[2])
+			idx := strings.TrimLeft(m[1], "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz")
+			if v, ok := cc[idx]; ok && strings.HasPrefix(m[1], "CUDA") {
+				d += " (" + v + ")"
+			}
+			c.Devices = append(c.Devices, d)
+			if c.Backend == "" {
+				c.Backend = strings.TrimRight(m[1], "0123456789")
+			}
+			known = true
+		}
+		if noDevicesRE.MatchString(l) {
+			c.NoGPU, known = true, true
+		}
+	}
+	switch {
+	case c.OnTheCPU():
+		c.Backend = "CPU"
+	case c.Backend == "" && c.Offloaded > 0:
+		c.Backend = "GPU"
+	case c.Backend == "":
+		c.Backend = "CPU"
+	}
+	return c, known
+}
+
+// OnTheCPU is whether a model the engine was given GPUs for is computed on
+// the CPU, by its own account: no device found, or no layer offloaded.
+func (c Compute) OnTheCPU() bool {
+	return c.NoGPU || (c.Total > 0 && c.Offloaded == 0)
+}

internal/agent/engines/compute_test.go

@@ -0,0 +1,32 @@
+package engines
+
+import (
+	"strings"
+	"testing"
+)
+
+func TestLlamaCppSaysWhereItPutTheModel(t *testing.T) {
+	gpu := strings.Split(`ggml_cuda_init: found 1 CUDA devices:
+  Device 0: Quadro RTX 6000, compute capability 7.5, VMM: yes
+load_backend: loaded CUDA backend from /opt/llmgw/engines/llamacpp/b10991/libggml-cuda.so
+llama_model_load_from_file_impl: using device CUDA0 (Quadro RTX 6000) - 23800 MiB free
+load_tensors: offloading 48 repeating layers to GPU
+load_tensors: offloaded 49/49 layers to GPU`, "\n")
+	c, ok := ReadLlamaCompute(gpu)
+	if !ok || c.Backend != "CUDA" || c.Offloaded != 49 || c.Total != 49 || c.OnTheCPU() ||
+		len(c.Devices) != 1 || c.Devices[0] != "CUDA0: Quadro RTX 6000 (7.5)" {
+		t.Errorf("on the GPU: %+v %v", c, ok)
+	}
+
+	cpu := strings.Split(`warning: no usable GPU found, --gpu-layers option will be ignored
+warning: one possible reason is that llama.cpp was compiled without GPU support
+load_tensors: offloaded 0/49 layers to GPU`, "\n")
+	c, ok = ReadLlamaCompute(cpu)
+	if !ok || c.Backend != "CPU" || !c.OnTheCPU() {
+		t.Errorf("fell back to the CPU: %+v %v", c, ok)
+	}
+
+	if _, ok := ReadLlamaCompute([]string{"main: server is listening on http://127.0.0.1:20001"}); ok {
+		t.Error("output that says nothing was read as saying something")
+	}
+}

internal/agent/engines/engine.go

@@ -114,6 +114,10 @@ type ShardSpec struct {
 	// lets the control plane tell, after its own restart, whether a shard that
 	// is still running matches the profile the catalog holds now.
 	ProfileFP string
+
+	// Compute is where the engine said it put the model as it became ready,
+	// kept with the shard's state so an agent restarted later still knows.
+	Compute *Compute `json:",omitempty"`
 }
 
 // Host is where the engine listens.

internal/agent/engines/llamacpp.go

@@ -277,6 +277,9 @@ func (l *LlamaCpp) bin() string {
 	return "llama-server"
 }
 
+// Binary is the llama-server this engine runs.
+func (l *LlamaCpp) Binary() string { return l.bin() }
+
 // FallsBackToCPU says that llama.cpp keeps running when its GPU backend does
 // not load: --n-gpu-layers asks for layers on a GPU it then has none of, and
 // the server comes up and answers on the CPU, at a tenth of the speed, with no

internal/agent/gpucheck.go

@@ -4,10 +4,12 @@ import (
 	"context"
 	"fmt"
 	"os"
+	"slices"
 	"strconv"
 	"strings"
 	"time"
 
+	llmgwv1 "gitea.mynodes.xyz/ilker/llmgw/api/gen/llmgw/v1"
 	"gitea.mynodes.xyz/ilker/llmgw/internal/agent/discovery"
 	"gitea.mynodes.xyz/ilker/llmgw/internal/agent/engines"
 )
@@ -153,3 +155,70 @@ func joinIndices(is []int) string {
 	}
 	return strings.Join(s, ", ")
 }
+
+// engineCompute is where an engine said it put its model, from its output
+// so far — llama.cpp's, which says — and false for one that said nothing. A
+// shard given GPUs that its engine says it computes on the CPU is an error,
+// found this way even where nvidia-smi cannot be asked.
+func engineCompute(eng engines.Engine, gpus []int, lines []string) (engines.Compute, bool, error) {
+	if _, ok := eng.(engines.CPUFallback); !ok {
+		return engines.Compute{}, false, nil
+	}
+	read, ok := engines.ReadLlamaCompute(lines)
+	if !ok {
+		return engines.Compute{}, false, nil
+	}
+	if len(gpus) > 0 && read.OnTheCPU() {
+		why := "its GPU backend saw no device it could use"
+		if read.Total > 0 && !read.NoGPU {
+			why = fmt.Sprintf("it offloaded %d of %d layers", read.Offloaded, read.Total)
+		}
+		return read, true, fmt.Errorf("placed on GPU %v but computing on the CPU: %s, by llama.cpp's own account "+
+			"(llmgw ctl engine-check on this node says why)", gpus, why)
+	}
+	return read, true, nil
+}
+
+// computeToWire is a compute as it travels to the control plane.
+func computeToWire(c *engines.Compute) *llmgwv1.Compute {
+	if c == nil {
+		return nil
+	}
+	return &llmgwv1.Compute{Backend: c.Backend, Devices: c.Devices,
+		LayersOffloaded: int32(c.Offloaded), LayersTotal: int32(c.Total)} //nolint:gosec // layer counts
+}
+
+// gpuCompute is where a shard computes as nvidia-smi sees it: the CPU for a
+// shard given no GPU, else the GPUs whose memory its processes hold, and how
+// much. False when nvidia-smi cannot be asked. Engine-agnostic, and needing
+// no log line: llama.cpp says where its layers went only at -lv 4.
+func (a *Agent) gpuCompute(ctx context.Context, gpus []int, pid int) (engines.Compute, bool) {
+	if len(gpus) == 0 {
+		return engines.Compute{Backend: "CPU"}, true
+	}
+	mem, err := gpuMemory(ctx, a.cfg.Discovery.NvidiaSMI)
+	if err != nil || mem == nil || pid <= 0 {
+		return engines.Compute{}, false
+	}
+	group := processGroup(pid)
+	c := engines.Compute{Backend: "CUDA"}
+	for _, g := range mem {
+		if !slices.Contains(gpus, g.Index) {
+			continue
+		}
+		used := 0
+		for _, p := range g.Procs {
+			if holds([]int{p.PID}, pid, group) {
+				used += p.UsedMB
+			}
+		}
+		if used > 0 {
+			c.Devices = append(c.Devices, fmt.Sprintf("GPU %d: %.1f GB", g.Index, float64(used)/1024))
+		}
+	}
+	if len(c.Devices) == 0 {
+		// Placed on GPUs and holding none of their memory.
+		c.Backend = "CPU"
+	}
+	return c, true
+}

internal/agent/gpucheck_test.go

@@ -122,3 +122,55 @@ func TestForeignGPUMemoryLeavesOutTheAgentsOwnShards(t *testing.T) {
 		t.Errorf("no reading came back as %v", foreign)
 	}
 }
+
+func TestLlamaCppsOwnWordOnTheCPUIsEnoughToRefuse(t *testing.T) {
+	eng := fallbackEngine{}
+	// Without nvidia-smi to ask, the engine's own output says it.
+	c, said, err := engineCompute(eng, []int{0}, []string{
+		"warning: no usable GPU found, --gpu-layers option will be ignored",
+		"load_tensors: offloaded 0/49 layers to GPU"})
+	if err == nil || !said || !strings.Contains(err.Error(), "computing on the CPU") || c.Backend != "CPU" {
+		t.Fatalf("a GPU shard llama.cpp put on the CPU: %v %+v", err, c)
+	}
+	c, said, err = engineCompute(eng, []int{0}, []string{
+		"llama_model_load_from_file_impl: using device CUDA0 (Quadro RTX 6000) - 23800 MiB free",
+		"load_tensors: offloaded 49/49 layers to GPU"})
+	if err != nil || !said || c.Backend != "CUDA" || c.Offloaded != 49 || computeToWire(&c).GetLayersOffloaded() != 49 {
+		t.Errorf("on its GPU: %v %+v", err, c)
+	}
+	// A CPU shard is not refused for being on the CPU, and says so.
+	if c, _, err := engineCompute(eng, nil, []string{"load_tensors: offloaded 0/49 layers to GPU"}); err != nil || c.Backend != "CPU" {
+		t.Errorf("a CPU shard: %v %+v", err, c)
+	}
+	// An engine that said nothing says nothing.
+	if _, said, err := engineCompute(eng, []int{0}, []string{"listening"}); err != nil || said {
+		t.Errorf("nothing said: %v", err)
+	}
+}
+
+func TestNvidiaSMISaysWhereAShardComputes(t *testing.T) {
+	oldMem, oldGroup := gpuMemory, processGroup
+	t.Cleanup(func() { gpuMemory, processGroup = oldMem, oldGroup })
+	processGroup = func(pid int) int { return pid }
+	gpuMemory = func(context.Context, string) ([]discovery.GPUMemory, error) {
+		return []discovery.GPUMemory{
+			{Index: 0, UsedMB: 20000, Procs: []discovery.GPUProcess{{PID: 700, UsedMB: 2048}, {PID: 900, UsedMB: 17000}}},
+			{Index: 1, UsedMB: 0},
+		}, nil
+	}
+	a := &Agent{log: slog.Default()}
+	if c, ok := a.gpuCompute(context.Background(), []int{0}, 700); !ok || c.Backend != "CUDA" ||
+		len(c.Devices) != 1 || c.Devices[0] != "GPU 0: 2.0 GB" {
+		t.Errorf("on GPU 0: %+v %v", c, ok)
+	}
+	if c, ok := a.gpuCompute(context.Background(), []int{1}, 700); !ok || c.Backend != "CPU" {
+		t.Errorf("placed on GPU 1, holding none of it: %+v", c)
+	}
+	if c, ok := a.gpuCompute(context.Background(), nil, 700); !ok || c.Backend != "CPU" {
+		t.Errorf("a CPU shard: %+v", c)
+	}
+	gpuMemory = func(context.Context, string) ([]discovery.GPUMemory, error) { return nil, nil }
+	if _, ok := a.gpuCompute(context.Background(), []int{0}, 700); ok {
+		t.Error("no nvidia-smi read as an answer")
+	}
+}

internal/agent/settings.go

@@ -0,0 +1,109 @@
+package agent
+
+import (
+	"context"
+	"crypto/sha256"
+	"os"
+	"path/filepath"
+	"time"
+
+	llmgwv1 "gitea.mynodes.xyz/ilker/llmgw/api/gen/llmgw/v1"
+)
+
+// The cluster's settings for this agent (BACKLOG E3): what agent.yaml can
+// say but the agent's id, control planes, address and certificate, kept by
+// the control planes and sent over the stream. Kept here beside agent.yaml,
+// which then holds only what the agent needs to start; at start the two are
+// read together, these over the file.
+
+// SettingsFile keeps the settings the control plane last sent.
+var SettingsFile = "/var/lib/llmgw-agent/agent-settings.yaml"
+
+// ReadSettings is what SettingsFile holds, nil when nothing.
+func ReadSettings() []byte {
+	b, err := os.ReadFile(SettingsFile)
+	if err != nil {
+		return nil
+	}
+	return b
+}
+
+// SetSettingsCheck says how to tell whether settings would load beside this
+// agent's own file; settings that would not are not taken.
+func (a *Agent) SetSettingsCheck(f func([]byte) error) { a.settingsCheck = f }
+
+// applySettings takes the settings the control plane sent: nothing when they
+// are what this agent runs on, and otherwise kept and started again on.
+func (a *Agent) applySettings(m *llmgwv1.AgentSettings) {
+	next := m.GetYaml()
+	if next == a.cfg.Settings {
+		return
+	}
+	if a.settingsCheck != nil {
+		if err := a.settingsCheck([]byte(next)); err != nil {
+			// Kept running on what it has: settings that will not load
+			// here must not take the node out of service.
+			a.log.Error("the cluster's settings for this agent do not load beside its agent.yaml; running on what it has",
+				"err", err)
+			return
+		}
+	}
+	if err := os.MkdirAll(filepath.Dir(SettingsFile), 0o755); err != nil {
+		a.log.Error("could not keep the cluster's settings for this agent", "err", err)
+		return
+	}
+	tmp := SettingsFile + ".new"
+	if err := os.WriteFile(tmp, []byte(next), 0o600); err != nil {
+		a.log.Error("could not keep the cluster's settings for this agent", "err", err)
+		return
+	}
+	if err := os.Rename(tmp, SettingsFile); err != nil {
+		a.log.Error("could not keep the cluster's settings for this agent", "err", err)
+		return
+	}
+	a.log.Info("the cluster's settings for this agent changed; starting again on them, the engines keep serving")
+	time.Sleep(300 * time.Millisecond)
+	exitForUpgrade()
+}
+
+// WatchFile starts the agent again when its configuration file changes and
+// still loads: agent.yaml is read only at start, and a change to it that
+// waited for somebody to restart the agent was a change nobody saw take
+// effect.
+func (a *Agent) WatchFile(ctx context.Context, path string, loads func() error) {
+	if path == "" {
+		return
+	}
+	sum := func() [32]byte {
+		b, _ := os.ReadFile(path) //nolint:gosec // the agent's own configuration
+		return sha256.Sum256(b)
+	}
+	was := sum()
+	t := time.NewTicker(watchEvery)
+	defer t.Stop()
+	for {
+		select {
+		case <-ctx.Done():
+			return
+		case <-t.C:
+		}
+		now := sum()
+		if now == was {
+			continue
+		}
+		if err := loads(); err != nil {
+			// Once per change: an editor saving halfway must not stop the
+			// agent, and a broken file is reported, not restarted into.
+			a.log.Error("agent.yaml changed and does not load; running on what it had", "path", path, "err", err)
+			was = now
+			continue
+		}
+		a.log.Info("agent.yaml changed; starting again on it, the engines keep serving", "path", path)
+		time.Sleep(300 * time.Millisecond)
+		exitForUpgrade()
+		return
+	}
+}
+
+// watchEvery is how often agent.yaml is looked at.
+var watchEvery = 10 * time.Second

internal/agent/settings_test.go

@@ -0,0 +1,88 @@
+package agent
+
+import (
+	"errors"
+	"os"
+	"path/filepath"
+	"sync/atomic"
+	"testing"
+	"time"
+
+	llmgwv1 "gitea.mynodes.xyz/ilker/llmgw/api/gen/llmgw/v1"
+)
+
+func settingsAgent(t *testing.T, running string) (*Agent, chan struct{}) {
+	t.Helper()
+	oldFile, oldExit := SettingsFile, exitForUpgrade
+	SettingsFile = filepath.Join(t.TempDir(), "agent-settings.yaml")
+	exited := make(chan struct{}, 1)
+	exitForUpgrade = func() { exited <- struct{}{} }
+	t.Cleanup(func() { SettingsFile, exitForUpgrade = oldFile, oldExit })
+	cfg := Config{Server: "127.0.0.1:7443", NodeID: "n1", WeightsCache: t.TempDir(), EnginesDir: t.TempDir(), Settings: running}
+	cfg.Superv.WorkDir = t.TempDir()
+	a, err := New(cfg, nil, nil)
+	if err != nil {
+		t.Fatal(err)
+	}
+	return a, exited
+}
+
+// exitedSoon is whether the agent exited within a second: it waits a
+// moment for its log line to leave first.
+func exitedSoon(ch chan struct{}) bool {
+	select {
+	case <-ch:
+		return true
+	case <-time.After(time.Second):
+		return false
+	}
+}
+
+func TestAnAgentOnItsSettingsStaysAndOneOffThemStartsAgain(t *testing.T) {
+	a, exited := settingsAgent(t, "heartbeat_interval: 5s\n")
+	a.applySettings(&llmgwv1.AgentSettings{Yaml: "heartbeat_interval: 5s\n"})
+	if exitedSoon(exited) || ReadSettings() != nil {
+		t.Error("settings it already runs on restarted it")
+	}
+	a.applySettings(&llmgwv1.AgentSettings{Yaml: "heartbeat_interval: 2s\n"})
+	if !exitedSoon(exited) || string(ReadSettings()) != "heartbeat_interval: 2s\n" {
+		t.Errorf("new settings: kept %q", ReadSettings())
+	}
+}
+
+func TestSettingsThatWouldNotLoadAreNotTaken(t *testing.T) {
+	a, exited := settingsAgent(t, "")
+	a.SetSettingsCheck(func([]byte) error { return errors.New("fabric ib-a lists no HCAs") })
+	a.applySettings(&llmgwv1.AgentSettings{Yaml: "fabrics: [{name: ib-a}]\n"})
+	if exitedSoon(exited) || ReadSettings() != nil {
+		t.Error("settings that do not load were taken")
+	}
+}
+
+func TestAChangedAgentYamlThatLoadsStartsTheAgentAgain(t *testing.T) {
+	a, exited := settingsAgent(t, "")
+	old := watchEvery
+	watchEvery = 10 * time.Millisecond
+	t.Cleanup(func() { watchEvery = old })
+	p := filepath.Join(t.TempDir(), "agent.yaml")
+	_ = os.WriteFile(p, []byte("node_id: n1\n"), 0o600)
+	var loads atomic.Bool
+	check := func() error {
+		if loads.Load() {
+			return nil
+		}
+		return errors.New("broken")
+	}
+	go a.WatchFile(t.Context(), p, check)
+	time.Sleep(30 * time.Millisecond)
+	// Saved broken: reported, not restarted into.
+	_ = os.WriteFile(p, []byte("node_id: [\n"), 0o600)
+	if exitedSoon(exited) {
+		t.Fatal("a broken agent.yaml restarted the agent")
+	}
+	loads.Store(true)
+	_ = os.WriteFile(p, []byte("node_id: n1\nlabels: {a: b}\n"), 0o600)
+	if !exitedSoon(exited) {
+		t.Error("a changed agent.yaml did not restart the agent")
+	}
+}

internal/agent/supervisor/runas.go

@@ -3,6 +3,7 @@ package supervisor
 import (
 	"fmt"
 	"os"
+	"os/exec"
 	"os/user"
 	"path/filepath"
 	"strconv"
@@ -90,3 +91,25 @@ func credentialFor(u *user.User, workDir string) (*runAs, error) {
 func (r *runAs) env() []string {
 	return []string{"HOME=" + r.home, "XDG_CACHE_HOME=" + filepath.Join(r.home, ".cache"), "USER=" + r.name, "LOGNAME=" + r.name}
 }
+
+// AsEngineUser has cmd run as the engine user, as an engine would, with its
+// home under workDir: for work the agent does that is not an engine but
+// reads what an engine reads — training an adapter. It answers that home
+// and a way to hand a path to the user, or why it runs as this process.
+func AsEngineUser(cmd *exec.Cmd, name, workDir string) (home string, give func(path string) error, why string) {
+	r, why := resolveRunAs(name, workDir)
+	if r == nil {
+		return "", func(string) error { return nil }, why
+	}
+	cmd.SysProcAttr = &syscall.SysProcAttr{Credential: r.cred, Setpgid: true}
+	cmd.Env = append(cmd.Env, r.env()...)
+	uid, gid := int(r.cred.Uid), int(r.cred.Gid)
+	return r.home, func(path string) error {
+		return filepath.WalkDir(path, func(p string, _ os.DirEntry, err error) error {
+			if err != nil {
+				return err
+			}
+			return os.Lchown(p, uid, gid)
+		})
+	}, ""
+}

internal/agent/supervisor/supervisor.go

@@ -113,7 +113,10 @@ type Shard struct {
 	exitCode int
 
 	logs *ringBuffer
-	cmd  *exec.Cmd
+	// cmdline is an adopted engine's command line, from its state file: it
+	// has no cmd of this agent's to read it from.
+	cmdline []string
+	cmd     *exec.Cmd
 	// sidecars are the shard's supporting processes, torn down with it.
 	sidecars []*exec.Cmd
 	cancel   context.CancelFunc
@@ -644,6 +647,32 @@ func (s *Supervisor) Get(id string) (*Shard, error) {
 	return sh, nil
 }
 
+// SetCompute records where a shard's engine said it put the model, with the
+// shard's state on disk.
+func (s *Supervisor) SetCompute(id string, c engines.Compute) {
+	s.mu.Lock()
+	sh, ok := s.shards[id]
+	if ok {
+		sh.Spec.Compute = &c
+	}
+	s.mu.Unlock()
+	if ok && sh.pid > 0 {
+		_ = s.writeState(sh)
+	}
+}
+
+// ComputeOf is where a shard's engine was found to compute, nil when not
+// yet known.
+func (s *Supervisor) ComputeOf(id string) *engines.Compute {
+	s.mu.Lock()
+	defer s.mu.Unlock()
+	if sh, ok := s.shards[id]; ok && sh.Spec.Compute != nil {
+		c := *sh.Spec.Compute
+		return &c
+	}
+	return nil
+}
+
 // Remove forgets a terminated shard.
 func (s *Supervisor) Remove(id string) {
 	s.mu.Lock()
@@ -703,11 +732,17 @@ type shardState struct {
 
 // writeState records everything needed to recognise and describe the process later.
 func (s *Supervisor) writeState(sh *Shard) error {
+	// An adopted engine was started by an earlier agent: its command line
+	// is the one its state file kept.
+	cmdline := sh.cmdline
+	if sh.cmd != nil {
+		cmdline = sh.cmd.Args
+	}
 	b, err := json.Marshal(shardState{
 		ShardID:   sh.ID,
 		ReplicaID: sh.Spec.ReplicaID,
 		PID:       sh.pid,
-		Cmdline:   sh.cmd.Args,
+		Cmdline:   cmdline,
 		StartedAt: sh.started,
 		Spec:      sh.Spec,
 	})
@@ -804,6 +839,7 @@ func (s *Supervisor) Adopt(o Orphan) (*Shard, error) {
 		started: o.StartedAt,
 		done:    make(chan struct{}),
 		adopted: true,
+		cmdline: o.Cmdline,
 	}
 	s.shards[o.ShardID] = sh
 	s.mu.Unlock()

internal/agent/supervisor/supervisor_test.go

@@ -848,3 +848,60 @@ func eventually(t *testing.T, limit time.Duration, msg string, cond func() bool)
 		time.Sleep(10 * time.Millisecond)
 	}
 }
+
+func TestWhatAShardComputesOnSurvivesTheAgent(t *testing.T) {
+	s, _ := newSupervisor(t)
+	spec := engines.ShardSpec{ReplicaID: "rep-1", ShardID: "a", HTTPPort: 20100}
+	sh, err := s.Start(t.Context(), "a", spec, shellCommand("sleep 30"))
+	if err != nil {
+		t.Fatal(err)
+	}
+	t.Cleanup(func() { _ = s.Stop(context.Background(), "a") })
+	_ = sh
+	s.SetCompute("a", engines.Compute{Backend: "CUDA", Offloaded: 49, Total: 49})
+
+	s2, err := New(Config{WorkDir: s.cfg.WorkDir, StopTimeout: 2 * time.Second},
+		&scriptedHealth{}, slog.New(slog.DiscardHandler))
+	if err != nil {
+		t.Fatal(err)
+	}
+	orphans, err := s2.FindOrphans()
+	if err != nil || len(orphans) != 1 {
+		t.Fatalf("%v %d", err, len(orphans))
+	}
+	if c := orphans[0].Spec.Compute; c == nil || c.Backend != "CUDA" || c.Offloaded != 49 {
+		t.Errorf("not kept: %+v", c)
+	}
+}
+
+func TestAnAdoptedShardsComputeIsKeptWithoutCrashing(t *testing.T) {
+	s, _ := newSupervisor(t)
+	spec := engines.ShardSpec{ReplicaID: "rep-1", ShardID: "a", HTTPPort: 20100}
+	if _, err := s.Start(t.Context(), "a", spec, shellCommand("sleep 30")); err != nil {
+		t.Fatal(err)
+	}
+	// An agent restart: the engine is adopted by a new supervisor, which did
+	// not start it and has no command of its own for it.
+	s2, err := New(Config{WorkDir: s.cfg.WorkDir, StopTimeout: 2 * time.Second},
+		&scriptedHealth{}, slog.New(slog.DiscardHandler))
+	if err != nil {
+		t.Fatal(err)
+	}
+	orphans, _ := s2.FindOrphans()
+	if len(orphans) != 1 {
+		t.Fatalf("%d orphans", len(orphans))
+	}
+	if _, err := s2.Adopt(orphans[0]); err != nil {
+		t.Fatal(err)
+	}
+	t.Cleanup(func() { _ = s2.Stop(context.Background(), "a") })
+	s2.SetCompute("a", engines.Compute{Backend: "CPU"})
+
+	// And a third agent still finds it, with its command line and compute.
+	s3, _ := New(Config{WorkDir: s.cfg.WorkDir, StopTimeout: 2 * time.Second},
+		&scriptedHealth{}, slog.New(slog.DiscardHandler))
+	again, _ := s3.FindOrphans()
+	if len(again) != 1 || len(again[0].Cmdline) == 0 || again[0].Spec.Compute == nil {
+		t.Fatalf("lost: %+v", again)
+	}
+}

internal/agent/train/runner.go

@@ -0,0 +1,574 @@
+// Package train runs LoRA training jobs on a node (BACKLOG F4): with an
+// installed engine's Python — its torch and transformers, already matched to
+// the node's CUDA — and a small overlay of what training needs besides, so a
+// node that serves can train without a second multi-gigabyte environment.
+package train
+
+import (
+	"archive/tar"
+	"bufio"
+	"bytes"
+	"compress/gzip"
+	"context"
+	"crypto/sha256"
+	_ "embed"
+	"encoding/hex"
+	"encoding/json"
+	"errors"
+	"fmt"
+	"io"
+	"log/slog"
+	"net/http"
+	"os"
+	"os/exec"
+	"path/filepath"
+	"sort"
+	"strconv"
+	"strings"
+	"sync"
+	"syscall"
+	"time"
+
+	"gitea.mynodes.xyz/ilker/llmgw/internal/agent/supervisor"
+)
+
+//go:embed trainer.py
+var trainerPy []byte
+
+// overlayPackages are what training needs besides an engine's Python, pinned
+// and installed without their dependencies: those are the engine's own, and
+// letting an installer resolve them pulled a second torch.
+var overlayPackages = []string{"peft==0.21.2", "accelerate==1.15.0"}
+
+// Job is one training job, as the control plane sent it.
+type Job struct {
+	ID           string
+	BaseRepo     string
+	BaseRevision string
+	// Data is NDJSON of {"q", "a"}.
+	Data []byte
+	// Params is the trainer's settings as JSON.
+	Params       string
+	GPUMemoryMiB int
+	GGUF         bool
+}
+
+// Status is how a job is going.
+type Status struct {
+	State   string
+	Message string
+	Step    int
+	Steps   int
+	Loss    float64
+	PeakMiB int
+}
+
+// Env is where a node keeps what training needs.
+type Env struct {
+	// EnginesDir is where engines are installed (the agent's engines_dir).
+	EnginesDir string
+	// WorkDir is the trainer's own directory: its overlay, the llama.cpp
+	// conversion code, and each job's files while it runs.
+	WorkDir string
+	// RunAs is the user jobs run as: the engines' user.
+	RunAs string
+	Log   *slog.Logger
+	// Python, UV, ConvDir and GPUs replace what would be found, for tests.
+	Python, UV, ConvDir string
+	GPUs                func(ctx context.Context) ([]GPU, error)
+}
+
+// GPU is one card and how much of it is free.
+type GPU struct {
+	Index   int
+	FreeMiB int
+}
+
+// Runner runs one job at a time.
+type Runner struct {
+	env    Env
+	mu     sync.Mutex
+	cancel context.CancelFunc
+	job    string
+}
+
+// New is a runner.
+func New(env Env) *Runner {
+	if env.Log == nil {
+		env.Log = slog.New(slog.DiscardHandler)
+	}
+	return &Runner{env: env}
+}
+
+// Cancel stops a job if it is the one running.
+func (r *Runner) Cancel(id string) {
+	r.mu.Lock()
+	defer r.mu.Unlock()
+	if r.job == id && r.cancel != nil {
+		r.cancel()
+	}
+}
+
+// ErrBusy is a node already training.
+var ErrBusy = errors.New("this node is already training")
+
+// Run runs a job to its end, telling status as it goes and artifact each
+// piece of what it made. It answers the error it failed with.
+func (r *Runner) Run(ctx context.Context, job Job, status func(Status), artifact func(kind string, seq int, data []byte, last bool) error) error {
+	r.mu.Lock()
+	if r.job != "" {
+		r.mu.Unlock()
+		return ErrBusy
+	}
+	ctx, cancel := context.WithCancel(ctx)
+	r.job, r.cancel = job.ID, cancel
+	r.mu.Unlock()
+	defer func() {
+		cancel()
+		r.mu.Lock()
+		r.job, r.cancel = "", nil
+		r.mu.Unlock()
+	}()
+
+	status(Status{State: "preparing", Message: "preparing the trainer"})
+	py, err := r.python()
+	if err != nil {
+		return err
+	}
+	overlay, err := r.overlay(ctx, py)
+	if err != nil {
+		return fmt.Errorf("installing %s: %w", strings.Join(overlayPackages, ", "), err)
+	}
+	conv := ""
+	if job.GGUF {
+		if conv, err = r.conversion(ctx); err != nil {
+			return fmt.Errorf("fetching llama.cpp's conversion code: %w", err)
+		}
+	}
+	gpu, err := r.pickGPU(ctx, job.GPUMemoryMiB)
+	if err != nil {
+		return err
+	}
+
+	dir := filepath.Join(r.env.WorkDir, "jobs", job.ID)
+	if err := os.MkdirAll(dir, 0o755); err != nil {
+		return err
+	}
+	defer os.RemoveAll(dir)
+	out := filepath.Join(dir, "adapter")
+	var params map[string]any
+	if job.Params != "" {
+		if err := json.Unmarshal([]byte(job.Params), &params); err != nil {
+			return fmt.Errorf("the job's params: %w", err)
+		}
+	}
+	spec, _ := json.Marshal(map[string]any{"base_repo": job.BaseRepo, "base_revision": job.BaseRevision,
+		"data": filepath.Join(dir, "data.jsonl"), "out": out, "params": params})
+	for name, b := range map[string][]byte{"data.jsonl": job.Data, "job.json": spec, "trainer.py": trainerPy} {
+		if err := os.WriteFile(filepath.Join(dir, name), b, 0o644); err != nil {
+			return err
+		}
+	}
+
+	var baseDir string
+	err = r.python3(ctx, py, dir, []string{"PYTHONPATH=" + overlay, "CUDA_VISIBLE_DEVICES=" + strconv.Itoa(gpu)},
+		[]string{filepath.Join(dir, "trainer.py"), filepath.Join(dir, "job.json")},
+		func(line map[string]any) {
+			s := Status{State: str(line["state"]), Message: str(line["message"]), Step: num(line["step"]),
+				Steps: num(line["steps"]), PeakMiB: num(line["peak_mib"])}
+			if v, ok := line["loss"].(float64); ok {
+				s.Loss = v
+			}
+			if b := str(line["base_dir"]); b != "" {
+				baseDir = b
+			}
+			if s.State == "trained" {
+				s.State = "running"
+				s.Message = "trained; packing the adapter"
+			}
+			status(s)
+		})
+	if err != nil {
+		return err
+	}
+
+	if job.GGUF {
+		status(Status{State: "running", Message: "converting the adapter to GGUF for llama.cpp"})
+		gguf := filepath.Join(dir, "adapter.gguf")
+		err := r.python3(ctx, py, dir, []string{"PYTHONPATH=" + strings.Join([]string{filepath.Join(conv, "gguf-py"), conv, overlay}, ":")},
+			[]string{filepath.Join(conv, "convert_lora_to_gguf.py"), "--base", baseDir, "--outtype", "f16", "--outfile", gguf, out}, nil)
+		if err != nil {
+			return fmt.Errorf("converting to GGUF: %w", err)
+		}
+		if err := sendFile(gguf, "gguf", artifact); err != nil {
+			return err
+		}
+	}
+	var tgz bytes.Buffer
+	if err := tarGz(&tgz, out); err != nil {
+		return err
+	}
+	return sendBytes(tgz.Bytes(), "peft", artifact)
+}
+
+func str(v any) string {
+	s, _ := v.(string)
+	return s
+}
+
+func num(v any) int {
+	f, _ := v.(float64)
+	return int(f)
+}
+
+// python3 runs a Python script as the engines' user, passing each JSON line
+// it prints to line, and answers the last of what it wrote to stderr when it
+// fails.
+func (r *Runner) python3(ctx context.Context, py, dir string, env, args []string, line func(map[string]any)) error {
+	cmd := exec.CommandContext(ctx, py, args...)
+	cmd.Dir = dir
+	cmd.Env = append(os.Environ(), "PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True", "TOKENIZERS_PARALLELISM=false",
+		"HF_HUB_DISABLE_PROGRESS_BARS=1", "PYTHONUNBUFFERED=1")
+	cmd.Env = append(cmd.Env, env...)
+	home, give, why := supervisor.AsEngineUser(cmd, r.env.RunAs, r.env.WorkDir)
+	if why != "" {
+		r.env.Log.Warn(why)
+	}
+	if home != "" {
+		cmd.Env = append(cmd.Env, "HF_HOME="+filepath.Join(home, ".cache", "huggingface"))
+	} else {
+		cmd.Env = append(cmd.Env, "HF_HOME="+filepath.Join(r.env.WorkDir, "hf"))
+	}
+	if err := give(dir); err != nil {
+		return fmt.Errorf("handing %s to %s: %w", dir, r.env.RunAs, err)
+	}
+	// The whole process group goes when the job is cancelled: torch starts
+	// workers of its own.
+	if cmd.SysProcAttr == nil {
+		cmd.SysProcAttr = &syscall.SysProcAttr{Setpgid: true}
+	}
+	cmd.Cancel = func() error { return syscall.Kill(-cmd.Process.Pid, syscall.SIGKILL) }
+	stdout, err := cmd.StdoutPipe()
+	if err != nil {
+		return err
+	}
+	var stderr tailBuffer
+	cmd.Stderr = &stderr
+	if err := cmd.Start(); err != nil {
+		return err
+	}
+	sc := bufio.NewScanner(stdout)
+	sc.Buffer(make([]byte, 64<<10), 1<<20)
+	for sc.Scan() {
+		var m map[string]any
+		if line != nil && json.Unmarshal(sc.Bytes(), &m) == nil {
+			line(m)
+		}
+	}
+	err = cmd.Wait()
+	if ctx.Err() != nil {
+		return errors.New("cancelled")
+	}
+	if err != nil {
+		return fmt.Errorf("%v: %s", err, stderr.last())
+	}
+	return nil
+}
+
+// tailBuffer keeps the end of what a process wrote: the reason it failed is
+// at the end, and the rest is progress bars.
+type tailBuffer struct{ b []byte }
+
+func (t *tailBuffer) Write(p []byte) (int, error) {
+	t.b = append(t.b, p...)
+	if len(t.b) > 16<<10 {
+		t.b = t.b[len(t.b)-16<<10:]
+	}
+	return len(p), nil
+}
+
+func (t *tailBuffer) last() string {
+	lines := strings.Split(strings.TrimSpace(string(t.b)), "\n")
+	if len(lines) > 3 {
+		lines = lines[len(lines)-3:]
+	}
+	return strings.Join(lines, " | ")
+}
+
+// python is the newest installed vLLM's interpreter: its torch is built for
+// this node's CUDA.
+func (r *Runner) python() (string, error) {
+	if r.env.Python != "" {
+		return r.env.Python, nil
+	}
+	found, _ := filepath.Glob(filepath.Join(r.env.EnginesDir, "vllm", "*", "bin", "python3"))
+	if len(found) == 0 {
+		return "", errors.New("training needs vLLM installed on this node (its torch is what trains); " +
+			"llmgw ctl engine-build, or serving any vLLM model here, installs it")
+	}
+	sort.Strings(found)
+	return found[len(found)-1], nil
+}
+
+func (r *Runner) uv() (string, error) {
+	if r.env.UV != "" {
+		return r.env.UV, nil
+	}
+	var found []string
+	_ = filepath.WalkDir(filepath.Join(r.env.EnginesDir, "uv"), func(p string, d os.DirEntry, err error) error {
+		if err == nil && !d.IsDir() && d.Name() == "uv" {
+			found = append(found, p)
+		}
+		return nil
+	})
+	if len(found) == 0 {
+		return "", errors.New("no uv under " + filepath.Join(r.env.EnginesDir, "uv"))
+	}
+	sort.Strings(found)
+	return found[len(found)-1], nil
+}
+
+// overlay installs the pinned packages once per Python and set of pins.
+func (r *Runner) overlay(ctx context.Context, py string) (string, error) {
+	sum := sha256.Sum256([]byte(py + "\x00" + strings.Join(overlayPackages, " ")))
+	dir := filepath.Join(r.env.WorkDir, "overlay-"+hex.EncodeToString(sum[:6]))
+	if _, err := os.Stat(filepath.Join(dir, ".done")); err == nil {
+		return dir, nil
+	}
+	uv, err := r.uv()
+	if err != nil {
+		return "", err
+	}
+	tmp := dir + ".new"
+	_ = os.RemoveAll(tmp)
+	args := append([]string{"pip", "install", "--quiet", "--python", py, "--no-deps", "--target", tmp}, overlayPackages...)
+	cmd := exec.CommandContext(ctx, uv, args...)
+	if out, err := cmd.CombinedOutput(); err != nil {
+		return "", fmt.Errorf("%v: %s", err, strings.TrimSpace(string(out)))
+	}
+	_ = os.WriteFile(filepath.Join(tmp, ".done"), nil, 0o644)
+	_ = os.RemoveAll(dir)
+	if err := os.Rename(tmp, dir); err != nil {
+		return "", err
+	}
+	// Readable by the engines' user, which runs the trainer.
+	_ = filepath.WalkDir(dir, func(p string, d os.DirEntry, err error) error {
+		if err == nil {
+			mode := os.FileMode(0o644)
+			if d.IsDir() {
+				mode = 0o755
+			}
+			_ = os.Chmod(p, mode)
+		}
+		return nil
+	})
+	return dir, nil
+}
+
+// llamaCppTag is the release of the newest llama.cpp installed: its
+// conversion code is the one whose GGUF that llama.cpp reads.
+func (r *Runner) llamaCppTag() (string, error) {
+	found, _ := filepath.Glob(filepath.Join(r.env.EnginesDir, "llamacpp", "b*"))
+	if len(found) == 0 {
+		return "", errors.New("a GGUF adapter needs llama.cpp installed on this node, to match its release")
+	}
+	sort.Strings(found)
+	tag, _, _ := strings.Cut(filepath.Base(found[len(found)-1]), "-")
+	return tag, nil
+}
+
+// conversionSource is where llama.cpp's source is fetched from, by tag.
+var conversionSource = "https://codeload.github.com/ggml-org/llama.cpp/tar.gz/refs/tags/"
+
+// conversion fetches llama.cpp's LoRA conversion code for the installed
+// release, once.
+func (r *Runner) conversion(ctx context.Context) (string, error) {
+	if r.env.ConvDir != "" {
+		return r.env.ConvDir, nil
+	}
+	tag, err := r.llamaCppTag()
+	if err != nil {
+		return "", err
+	}
+	dir := filepath.Join(r.env.WorkDir, "llama.cpp-"+tag)
+	if _, err := os.Stat(filepath.Join(dir, "convert_lora_to_gguf.py")); err == nil {
+		return dir, nil
+	}
+	req, err := http.NewRequestWithContext(ctx, http.MethodGet, conversionSource+tag, nil)
+	if err != nil {
+		return "", err
+	}
+	res, err := http.DefaultClient.Do(req)
+	if err != nil {
+		return "", err
+	}
+	defer res.Body.Close()
+	if res.StatusCode != http.StatusOK {
+		return "", fmt.Errorf("%s answered %s", conversionSource+tag, res.Status)
+	}
+	tmp := dir + ".new"
+	_ = os.RemoveAll(tmp)
+	if err := untarConversion(res.Body, tmp); err != nil {
+		return "", err
+	}
+	_ = os.RemoveAll(dir)
+	return dir, os.Rename(tmp, dir)
+}
+
+// untarConversion keeps, of llama.cpp's source, the conversion scripts and
+// the Python packages they import.
+func untarConversion(r io.Reader, dir string) error {
+	gz, err := gzip.NewReader(r)
+	if err != nil {
+		return err
+	}
+	tr := tar.NewReader(gz)
+	kept := 0
+	for {
+		h, err := tr.Next()
+		if errors.Is(err, io.EOF) {
+			break
+		}
+		if err != nil {
+			return err
+		}
+		_, rel, ok := strings.Cut(h.Name, "/")
+		if !ok || h.Typeflag != tar.TypeReg || strings.Contains(rel, "..") {
+			continue
+		}
+		if rel != "convert_lora_to_gguf.py" && rel != "convert_hf_to_gguf.py" &&
+			!strings.HasPrefix(rel, "conversion/") && !strings.HasPrefix(rel, "gguf-py/") {
+			continue
+		}
+		p := filepath.Join(dir, filepath.FromSlash(rel))
+		if err := os.MkdirAll(filepath.Dir(p), 0o755); err != nil {
+			return err
+		}
+		f, err := os.OpenFile(p, os.O_CREATE|os.O_WRONLY|os.O_TRUNC, 0o644)
+		if err != nil {
+			return err
+		}
+		_, err = io.Copy(f, io.LimitReader(tr, 64<<20))
+		f.Close()
+		if err != nil {
+			return err
+		}
+		kept++
+	}
+	if kept == 0 {
+		return errors.New("no conversion code in llama.cpp's source")
+	}
+	return nil
+}
+
+// pickGPU is the card with the most memory free, if that is enough.
+func (r *Runner) pickGPU(ctx context.Context, need int) (int, error) {
+	list := r.env.GPUs
+	if list == nil {
+		list = nvidiaSMI
+	}
+	gpus, err := list(ctx)
+	if err != nil {
+		return 0, fmt.Errorf("reading the cards' free memory: %w", err)
+	}
+	if len(gpus) == 0 {
+		return 0, errors.New("this node has no GPU to train on")
+	}
+	best := gpus[0]
+	for _, g := range gpus[1:] {
+		if g.FreeMiB > best.FreeMiB {
+			best = g
+		}
+	}
+	if best.FreeMiB < need {
+		return 0, fmt.Errorf("no card here has %d MiB free (the most free is GPU %d with %d MiB); "+
+			"a smaller base, or a card freed for it, would do", need, best.Index, best.FreeMiB)
+	}
+	return best.Index, nil
+}
+
+func nvidiaSMI(ctx context.Context) ([]GPU, error) {
+	ctx, cancel := context.WithTimeout(ctx, 15*time.Second)
+	defer cancel()
+	out, err := exec.CommandContext(ctx, "nvidia-smi", "--query-gpu=index,memory.free", "--format=csv,noheader,nounits").Output()
+	if err != nil {
+		return nil, err
+	}
+	return parseGPUs(string(out)), nil
+}
+
+func parseGPUs(s string) []GPU {
+	var out []GPU
+	for _, line := range strings.Split(strings.TrimSpace(s), "\n") {
+		a, b, ok := strings.Cut(line, ",")
+		if !ok {
+			continue
+		}
+		i, err1 := strconv.Atoi(strings.TrimSpace(a))
+		f, err2 := strconv.Atoi(strings.TrimSpace(b))
+		if err1 == nil && err2 == nil {
+			out = append(out, GPU{Index: i, FreeMiB: f})
+		}
+	}
+	return out
+}
+
+// tarGz packs a directory.
+func tarGz(w io.Writer, dir string) error {
+	gz := gzip.NewWriter(w)
+	tw := tar.NewWriter(gz)
+	err := filepath.WalkDir(dir, func(p string, d os.DirEntry, err error) error {
+		if err != nil || d.IsDir() {
+			return err
+		}
+		rel, _ := filepath.Rel(dir, p)
+		info, err := d.Info()
+		if err != nil {
+			return err
+		}
+		if err := tw.WriteHeader(&tar.Header{Name: filepath.ToSlash(rel), Mode: 0o644, Size: info.Size(), ModTime: info.ModTime()}); err != nil {
+			return err
+		}
+		f, err := os.Open(p)
+		if err != nil {
+			return err
+		}
+		defer f.Close()
+		_, err = io.Copy(tw, f)
+		return err
+	})
+	if err != nil {
+		return err
+	}
+	if err := tw.Close(); err != nil {
+		return err
+	}
+	return gz.Close()
+}
+
+// artifactChunk is how much of an artifact goes in one message.
+const artifactChunk = 1 << 20
+
+func sendFile(path, kind string, send func(string, int, []byte, bool) error) error {
+	b, err := os.ReadFile(path)
+	if err != nil {
+		return err
+	}
+	return sendBytes(b, kind, send)
+}
+
+func sendBytes(b []byte, kind string, send func(string, int, []byte, bool) error) error {
+	seq := 0
+	for {
+		n := min(artifactChunk, len(b))
+		last := n == len(b)
+		if err := send(kind, seq, b[:n], last); err != nil {
+			return err
+		}
+		if last {
+			return nil
+		}
+		b = b[n:]
+		seq++
+	}
+}

internal/agent/train/runner_test.go

@@ -0,0 +1,164 @@
+package train
+
+import (
+	"archive/tar"
+	"bytes"
+	"compress/gzip"
+	"context"
+	"io"
+	"os"
+	"path/filepath"
+	"strings"
+	"sync"
+	"testing"
+)
+
+// fakePython stands in for an engine's interpreter: as the trainer it
+// prints progress and writes an adapter; as the converter it writes a GGUF;
+// told to fail, it fails.
+const fakePython = `#!/bin/sh
+case "$1" in
+*trainer.py)
+  out=$(sed -n 's/.*"out": *"\([^"]*\)".*/\1/p' "$2")
+  grep -q FAIL "$(dirname "$2")/data.jsonl" && { echo "CUDA out of memory" >&2; exit 1; }
+  echo '{"state":"preparing","message":"fetching base"}'
+  echo '{"state":"running","step":5,"steps":10,"loss":0.9,"peak_mib":1900}'
+  mkdir -p "$out"; echo '{"r":16}' > "$out/adapter_config.json"; echo weights > "$out/adapter_model.safetensors"
+  echo '{"state":"trained","step":10,"steps":10,"peak_mib":2100,"base_dir":"/base"}'
+  ;;
+*convert_lora_to_gguf.py)
+  while [ $# -gt 0 ]; do [ "$1" = "--outfile" ] && echo GGUF > "$2"; shift; done
+  ;;
+esac
+`
+
+func runner(t *testing.T, free int) *Runner {
+	t.Helper()
+	dir := t.TempDir()
+	py := filepath.Join(dir, "python3")
+	_ = os.WriteFile(py, []byte(fakePython), 0o755)
+	uv := filepath.Join(dir, "uv")
+	_ = os.WriteFile(uv, []byte("#!/bin/sh\nwhile [ $# -gt 0 ]; do [ \"$1\" = \"--target\" ] && mkdir -p \"$2\"; shift; done\n"), 0o755)
+	return New(Env{WorkDir: filepath.Join(dir, "work"), Python: py, UV: uv, ConvDir: dir,
+		GPUs: func(context.Context) ([]GPU, error) { return []GPU{{0, 500}, {1, free}}, nil }})
+}
+
+type sink struct {
+	mu       sync.Mutex
+	statuses []Status
+	parts    map[string][]byte
+	lasts    int
+}
+
+func (s *sink) status(st Status) { s.mu.Lock(); s.statuses = append(s.statuses, st); s.mu.Unlock() }
+func (s *sink) artifact(kind string, _ int, b []byte, last bool) error {
+	s.mu.Lock()
+	defer s.mu.Unlock()
+	if s.parts == nil {
+		s.parts = map[string][]byte{}
+	}
+	s.parts[kind] = append(s.parts[kind], b...)
+	if last {
+		s.lasts++
+	}
+	return nil
+}
+
+func TestAJobTrainsOnTheFreestCardAndSendsWhatItMade(t *testing.T) {
+	r := runner(t, 3000)
+	s := &sink{}
+	err := r.Run(context.Background(), Job{ID: "j1", BaseRepo: "Qwen/x", Data: []byte(`{"q":"a","a":"b"}`), GPUMemoryMiB: 2500, GGUF: true},
+		s.status, s.artifact)
+	if err != nil {
+		t.Fatal(err)
+	}
+	var running bool
+	for _, st := range s.statuses {
+		if st.State == "running" && st.Step == 5 && st.Steps == 10 && st.PeakMiB == 1900 {
+			running = true
+		}
+	}
+	if !running {
+		t.Errorf("statuses %+v", s.statuses)
+	}
+	if string(s.parts["gguf"]) != "GGUF\n" || s.lasts != 2 {
+		t.Errorf("gguf %q, lasts %d", s.parts["gguf"], s.lasts)
+	}
+	files := map[string]string{}
+	gz, err := gzip.NewReader(bytes.NewReader(s.parts["peft"]))
+	if err != nil {
+		t.Fatal(err)
+	}
+	tr := tar.NewReader(gz)
+	for {
+		h, err := tr.Next()
+		if err != nil {
+			break
+		}
+		b, _ := io.ReadAll(tr)
+		files[h.Name] = string(b)
+	}
+	if files["adapter_model.safetensors"] != "weights\n" || files["adapter_config.json"] == "" {
+		t.Errorf("peft archive %v", files)
+	}
+	// Nothing of the job is left on the node.
+	if entries, _ := os.ReadDir(filepath.Join(r.env.WorkDir, "jobs")); len(entries) != 0 {
+		t.Errorf("left behind: %v", entries)
+	}
+}
+
+func TestAJobWaitsForNoCardItDoesNotFit(t *testing.T) {
+	r := runner(t, 1000)
+	err := r.Run(context.Background(), Job{ID: "j1", Data: []byte(`{"q":"a","a":"b"}`), GPUMemoryMiB: 2500}, (&sink{}).status, (&sink{}).artifact)
+	if err == nil || !strings.Contains(err.Error(), "no card here has 2500 MiB free") {
+		t.Errorf("got %v", err)
+	}
+}
+
+func TestATrainerThatFailsSaysWhy(t *testing.T) {
+	r := runner(t, 3000)
+	err := r.Run(context.Background(), Job{ID: "j1", Data: []byte(`{"q":"FAIL","a":"b"}`)}, (&sink{}).status, (&sink{}).artifact)
+	if err == nil || !strings.Contains(err.Error(), "CUDA out of memory") {
+		t.Errorf("got %v", err)
+	}
+}
+
+func TestOneJobAtATime(t *testing.T) {
+	r := runner(t, 3000)
+	r.job = "other"
+	if err := r.Run(context.Background(), Job{ID: "j2"}, (&sink{}).status, (&sink{}).artifact); err != ErrBusy {
+		t.Errorf("got %v", err)
+	}
+}
+
+func TestCardsAreReadFromNvidiaSMI(t *testing.T) {
+	got := parseGPUs("0, 2310\n1, 80000\nbad line\n")
+	if len(got) != 2 || got[1].FreeMiB != 80000 || got[0].Index != 0 {
+		t.Errorf("parsed %+v", got)
+	}
+}
+
+func TestOnlyTheConversionCodeIsKeptOfLlamaCpp(t *testing.T) {
+	var buf bytes.Buffer
+	gz := gzip.NewWriter(&buf)
+	tw := tar.NewWriter(gz)
+	for _, name := range []string{"llama.cpp-b1/convert_lora_to_gguf.py", "llama.cpp-b1/conversion/qwen.py",
+		"llama.cpp-b1/gguf-py/gguf/__init__.py", "llama.cpp-b1/src/llama.cpp", "llama.cpp-b1/../evil"} {
+		_ = tw.WriteHeader(&tar.Header{Name: name, Mode: 0o644, Size: 1, Typeflag: tar.TypeReg})
+		_, _ = tw.Write([]byte("x"))
+	}
+	_ = tw.Close()
+	_ = gz.Close()
+	dir := filepath.Join(t.TempDir(), "conv")
+	if err := untarConversion(&buf, dir); err != nil {
+		t.Fatal(err)
+	}
+	for _, want := range []string{"convert_lora_to_gguf.py", "conversion/qwen.py", "gguf-py/gguf/__init__.py"} {
+		if _, err := os.Stat(filepath.Join(dir, want)); err != nil {
+			t.Errorf("%s not kept", want)
+		}
+	}
+	if _, err := os.Stat(filepath.Join(dir, "src")); err == nil {
+		t.Error("the rest of llama.cpp was kept")
+	}
+}

internal/agent/train/trainer.py

@@ -0,0 +1,101 @@
+# llmgw's LoRA trainer (BACKLOG F4), run by the node agent with an installed
+# engine's Python (its torch and transformers) and a small overlay (peft,
+# accelerate). Reads a job, trains, writes the PEFT adapter, and says how it
+# goes as one JSON object a line on standard output.
+import json, math, os, random, sys, threading, subprocess, time
+
+def say(**kw):
+    print(json.dumps(kw), flush=True)
+
+job = json.load(open(sys.argv[1], encoding="utf-8"))
+p = {"rank": 16, "alpha": 32, "dropout": 0.05, "epochs": 6, "lr": 2e-4, "batch": 4, "max_len": 512, "seed": 0}
+p.update(job.get("params") or {})
+
+say(state="preparing", message="loading libraries")
+import torch
+from huggingface_hub import snapshot_download
+from transformers import AutoModelForCausalLM, AutoTokenizer
+from peft import LoraConfig, get_peft_model
+
+random.seed(p["seed"]); torch.manual_seed(p["seed"])
+if not torch.cuda.is_available():
+    raise SystemExit("no CUDA device is visible to the trainer")
+
+say(state="preparing", message="fetching " + job["base_repo"])
+base_dir = snapshot_download(job["base_repo"], revision=job.get("base_revision") or None,
+                             allow_patterns=["*.json", "*.safetensors", "*.txt", "*.model", "*.tiktoken"])
+say(state="preparing", message="loading the base model", base_dir=base_dir)
+
+# bf16 where the card has it; fp16 with a loss scaler where it does not
+# (Turing and older).
+bf16 = torch.cuda.is_bf16_supported()
+dtype = torch.bfloat16 if bf16 else torch.float16
+tok = AutoTokenizer.from_pretrained(base_dir)
+if tok.pad_token_id is None:
+    tok.pad_token = tok.eos_token
+model = AutoModelForCausalLM.from_pretrained(base_dir, dtype=dtype).to("cuda")
+
+rows = [json.loads(l) for l in open(job["data"], encoding="utf-8") if l.strip()]
+rows = [r for r in rows if r.get("q") and r.get("a")]
+if not rows:
+    raise SystemExit("no {q, a} pairs in the data")
+
+def encode(r):
+    prompt = tok.apply_chat_template([{"role": "user", "content": r["q"]}], add_generation_prompt=True, tokenize=False)
+    full = tok.apply_chat_template([{"role": "user", "content": r["q"]}, {"role": "assistant", "content": r["a"]}], tokenize=False)
+    pi = tok(prompt, add_special_tokens=False)["input_ids"]
+    fi = tok(full, add_special_tokens=False)["input_ids"][: p["max_len"]]
+    return fi, ([-100] * len(pi) + fi[len(pi):])[: len(fi)]
+
+data = [encode(r) for r in rows]
+
+model.gradient_checkpointing_enable()
+model.enable_input_require_grads()
+model = get_peft_model(model, LoraConfig(r=int(p["rank"]), lora_alpha=int(p["alpha"]), lora_dropout=float(p["dropout"]),
+    task_type="CAUSAL_LM", target_modules=["q_proj", "k_proj", "v_proj", "o_proj", "gate_proj", "up_proj", "down_proj"]))
+trainable = [x for x in model.parameters() if x.requires_grad]
+for x in trainable:  # fp32 trainable weights, so a loss scaler can unscale them
+    x.data = x.data.float()
+
+peak = {"mib": 0}
+stop = threading.Event()
+def watch():
+    while not stop.is_set():
+        try:
+            out = subprocess.run(["nvidia-smi", "--query-compute-apps=pid,used_memory", "--format=csv,noheader,nounits"],
+                                 capture_output=True, text=True, timeout=5).stdout
+            for line in out.splitlines():
+                pid, mib = [v.strip() for v in line.split(",")]
+                if int(pid) == os.getpid():
+                    peak["mib"] = max(peak["mib"], int(mib))
+        except Exception:
+            pass
+        time.sleep(0.5)
+threading.Thread(target=watch, daemon=True).start()
+
+opt = torch.optim.AdamW(trainable, lr=float(p["lr"]), weight_decay=0.0)
+scaler = torch.amp.GradScaler("cuda", enabled=not bf16)
+bs, epochs = int(p["batch"]), int(p["epochs"])
+steps = epochs * math.ceil(len(data) / bs)
+warm = max(1, steps // 10)
+sched = torch.optim.lr_scheduler.LambdaLR(opt, lambda s: min(1.0, (s + 1) / warm) * max(0.0, 1 - s / steps))
+model.train(); step = 0; last = time.time()
+for ep in range(epochs):
+    random.shuffle(data)
+    for i in range(0, len(data), bs):
+        batch = data[i:i + bs]; n = max(len(x[0]) for x in batch)
+        ids = torch.tensor([x[0] + [tok.pad_token_id] * (n - len(x[0])) for x in batch], device="cuda")
+        lab = torch.tensor([x[1] + [-100] * (n - len(x[1])) for x in batch], device="cuda")
+        att = torch.tensor([[1] * len(x[0]) + [0] * (n - len(x[0])) for x in batch], device="cuda")
+        with torch.autocast("cuda", dtype=dtype):
+            loss = model(input_ids=ids, attention_mask=att, labels=lab).loss
+        scaler.scale(loss).backward(); scaler.unscale_(opt)
+        torch.nn.utils.clip_grad_norm_(trainable, 1.0)
+        scaler.step(opt); scaler.update(); opt.zero_grad(set_to_none=True); sched.step(); step += 1
+        if time.time() - last > 2 or step == steps:
+            last = time.time()
+            say(state="running", message="training", step=step, steps=steps, loss=round(loss.item(), 4), peak_mib=peak["mib"])
+
+model.save_pretrained(job["out"])
+stop.set()
+say(state="trained", step=step, steps=steps, peak_mib=peak["mib"], base_dir=base_dir)

internal/agent/trainjob.go

@@ -0,0 +1,46 @@
+package agent
+
+import (
+	"context"
+
+	llmgwv1 "gitea.mynodes.xyz/ilker/llmgw/api/gen/llmgw/v1"
+	"gitea.mynodes.xyz/ilker/llmgw/internal/agent/train"
+)
+
+// trainJob runs a training job the control plane sent, telling it how the
+// job goes and sending what it made. A stream that ends ends the job: the
+// control plane, which sees the node go, marks it failed, and it is started
+// again from the beginning rather than resumed.
+func (a *Agent) trainJob(ctx context.Context, m *llmgwv1.TrainStart, out chan<- *llmgwv1.AgentMessage) {
+	id := m.GetJobId()
+	send := func(msg *llmgwv1.AgentMessage) bool {
+		select {
+		case out <- msg:
+			return true
+		case <-ctx.Done():
+			return false
+		}
+	}
+	status := func(s train.Status) {
+		send(&llmgwv1.AgentMessage{Msg: &llmgwv1.AgentMessage_TrainStatus{TrainStatus: &llmgwv1.TrainStatus{
+			JobId: id, State: s.State, Message: s.Message, Step: int32(s.Step), Steps: int32(s.Steps), //nolint:gosec // counts
+			Loss: s.Loss, PeakMib: int32(s.PeakMiB)}}}) //nolint:gosec // MiB
+	}
+	artifact := func(kind string, seq int, data []byte, last bool) error {
+		if !send(&llmgwv1.AgentMessage{Msg: &llmgwv1.AgentMessage_TrainArtifact{TrainArtifact: &llmgwv1.TrainArtifact{
+			JobId: id, Kind: kind, Seq: int32(seq), Data: data, Last: last}}}) { //nolint:gosec // chunk count
+			return ctx.Err()
+		}
+		return nil
+	}
+	a.log.Info("training an adapter", "job", id, "base", m.GetBaseRepo())
+	err := a.trainer.Run(ctx, train.Job{ID: id, BaseRepo: m.GetBaseRepo(), BaseRevision: m.GetBaseRevision(),
+		Data: m.GetData(), Params: m.GetParams(), GPUMemoryMiB: int(m.GetGpuMemoryMib()), GGUF: m.GetGguf()}, status, artifact)
+	if err != nil {
+		a.log.Warn("training failed", "job", id, "err", err)
+		status(train.Status{State: "failed", Message: err.Error()})
+		return
+	}
+	a.log.Info("trained an adapter", "job", id)
+	status(train.Status{State: "done"})
+}

internal/agent/watchexits_test.go

@@ -10,6 +10,7 @@ import (
 	llmgwv1 "gitea.mynodes.xyz/ilker/llmgw/api/gen/llmgw/v1"
 	"gitea.mynodes.xyz/ilker/llmgw/internal/agent/engines"
 	"gitea.mynodes.xyz/ilker/llmgw/internal/agent/supervisor"
+	"gitea.mynodes.xyz/ilker/llmgw/internal/control/agentsvc"
 )
 
 type alwaysReady struct{}
@@ -40,7 +41,7 @@ func TestAnAdoptedEngineThatDiesIsReported(t *testing.T) {
 	out := make(chan *llmgwv1.AgentMessage, 4)
 	ctx, cancel := context.WithCancel(t.Context())
 	defer cancel()
-	go a.watchExits(ctx, out)
+	go a.watchExits(ctx, out, nil)
 	time.Sleep(1500 * time.Millisecond) // seen ready
 	_ = cmd.Process.Kill()
 
@@ -54,3 +55,50 @@ func TestAnAdoptedEngineThatDiesIsReported(t *testing.T) {
 		t.Fatal("the adopted engine died and nothing was reported")
 	}
 }
+
+// Registered while its health check still ran: the ready that follows is
+// reported, or the control plane keeps the replica starting while it serves.
+func TestAShardRegisteredStartingIsReportedReady(t *testing.T) {
+	sup, err := supervisor.New(supervisor.Config{WorkDir: t.TempDir(), RunAs: "-"}, alwaysReady{}, slog.New(slog.DiscardHandler))
+	if err != nil {
+		t.Fatal(err)
+	}
+	cmd := exec.Command("sleep", "60")
+	if err := cmd.Start(); err != nil {
+		t.Fatal(err)
+	}
+	defer func() { _ = cmd.Process.Kill(); _ = cmd.Wait() }()
+	sh, err := sup.Adopt(supervisor.Orphan{ShardID: "m-r0", ReplicaID: "m", PID: cmd.Process.Pid,
+		Spec: engines.ShardSpec{ReplicaID: "m", ShardID: "m-r0", HTTPPort: 1}, Alive: true})
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	a := &Agent{sup: sup, log: slog.New(slog.DiscardHandler)}
+	pending := unready(a.runningShards())
+	if !pending["m-r0"] {
+		t.Fatalf("registration reported it ready already: %v", pending)
+	}
+	out := make(chan *llmgwv1.AgentMessage, 4)
+	ctx, cancel := context.WithCancel(t.Context())
+	defer cancel()
+	go a.watchExits(ctx, out, pending)
+	if err := sup.WaitReady(t.Context(), sh, "http://127.0.0.1:1", "m", 5*time.Second); err != nil {
+		t.Fatal(err)
+	}
+
+	select {
+	case m := <-out:
+		s := m.GetShard()
+		if s.GetShardId() != "m-r0" || s.GetState() != agentsvc.StateToWire(supervisor.Ready) {
+			t.Errorf("reported %+v", s)
+		}
+	case <-time.After(10 * time.Second):
+		t.Fatal("it became ready and nothing was reported")
+	}
+	select {
+	case m := <-out:
+		t.Errorf("reported twice: %+v", m.GetShard())
+	case <-time.After(2500 * time.Millisecond):
+	}
+}

internal/audit/audit.go

@@ -12,6 +12,7 @@ import (
 	"net/http"
 	"net/url"
 	"regexp"
+	"slices"
 	"sort"
 	"strings"
 	"time"
@@ -110,11 +111,16 @@ func (w *statusWriter) Flush() {
 	}
 }
 
+// notChanges are POSTs that change nothing and carry text that must not be
+// kept: asking a judge about a sentence.
+var notChanges = []string{"/api/v1/guard/judges/ask"}
+
 // API records every change to the admin API: a request that is not a GET or
 // HEAD, under /api/. actor says who made it.
 func API(sink Sink, actor func(*http.Request) string, log *slog.Logger, next http.Handler) http.Handler {
 	return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
-		if r.Method == http.MethodGet || r.Method == http.MethodHead || !strings.HasPrefix(r.URL.Path, "/api/") {
+		if r.Method == http.MethodGet || r.Method == http.MethodHead || !strings.HasPrefix(r.URL.Path, "/api/") ||
+			slices.Contains(notChanges, r.URL.Path) {
 			next.ServeHTTP(w, r)
 			return
 		}

internal/audit/audit_test.go

@@ -72,3 +72,13 @@ func TestABodyLargerThanWhatIsRecordedReachesTheHandlerWhole(t *testing.T) {
 		t.Errorf("the handler was given %d of %d bytes", got, len(big))
 	}
 }
+
+func TestAskingAJudgeKeepsNoText(t *testing.T) {
+	sink := &memSink{}
+	h := API(sink, func(*http.Request) string { return "ilker" }, nil, http.HandlerFunc(func(http.ResponseWriter, *http.Request) {}))
+	h.ServeHTTP(httptest.NewRecorder(), httptest.NewRequest("POST", "/api/v1/guard/judges/ask",
+		strings.NewReader(`{"text":"something a user wrote"}`)))
+	if len(sink.got) != 0 {
+		t.Errorf("a judge's question was recorded: %+v", sink.got)
+	}
+}

internal/cli/agent.go

@@ -40,10 +40,23 @@ func newAgentCmd() *cobra.Command {
 		RunE: func(cmd *cobra.Command, _ []string) error {
 			log := logger()
 
-			cfg, err := config.LoadAgent(configPath)
+			// The cluster's settings for this agent over its file, as the
+			// control plane last sent them. Settings that no longer load
+			// beside the file are set aside rather than keeping the agent
+			// from starting: the control plane sends them again, and says
+			// why they are refused.
+			settings := agent.ReadSettings()
+			cfg, err := config.LoadAgentWith(configPath, settings)
+			if err != nil && settings != nil {
+				log.Warn("the cluster's settings kept for this agent do not load beside its agent.yaml; starting without them",
+					"err", err)
+				settings = nil
+				cfg, err = config.LoadAgent(configPath)
+			}
 			if err != nil {
 				return err
 			}
+			cfg.Settings = string(settings)
 			if server != "" {
 				cfg.Server = server
 			}
@@ -89,9 +102,17 @@ func newAgentCmd() *cobra.Command {
 			if err != nil {
 				return err
 			}
+			a.SetSettingsCheck(func(b []byte) error {
+				_, err := config.LoadAgentWith(configPath, b)
+				return err
+			})
 
 			ctx, stop := signal.NotifyContext(cmd.Context(), syscall.SIGINT, syscall.SIGTERM)
 			defer stop()
+			go a.WatchFile(ctx, configPath, func() error {
+				_, err := config.LoadAgentWith(configPath, agent.ReadSettings())
+				return err
+			})
 
 			log.Info("agent starting",
 				"node", cfg.NodeID, "server", cfg.Server,

internal/cli/backup.go

@@ -249,7 +249,14 @@ func pgEnv(dsn string) ([]string, error) {
 		set("PGPASSWORD", pw)
 	}
 	set("PGDATABASE", strings.TrimPrefix(u.Path, "/"))
-	set("PGSSLMODE", u.Query().Get("sslmode"))
+	// The TLS settings go with the mode: verify-full without the cluster's
+	// authority sent libpq looking for ~/.postgresql/root.crt, and the dump
+	// failed on every control plane whose store is replicated.
+	q := u.Query()
+	set("PGSSLMODE", q.Get("sslmode"))
+	set("PGSSLROOTCERT", q.Get("sslrootcert"))
+	set("PGSSLCERT", q.Get("sslcert"))
+	set("PGSSLKEY", q.Get("sslkey"))
 	// A standby if there is one, else whichever answers — not the address's
 	// read-write: a dump only reads, it need not load the primary, and one
 	// that insists on the primary fails exactly while the cluster is failing

internal/cli/backup_test.go

@@ -238,3 +238,17 @@ func TestAStoreOfSeveralHostsIsGivenToPgDumpAsLibpqLists(t *testing.T) {
 		t.Error("an empty host in the list was accepted")
 	}
 }
+
+func TestPgDumpIsGivenTheClusterAuthorityAlongWithVerifyFull(t *testing.T) {
+	// Without it libpq looked for ~/.postgresql/root.crt and the dump failed.
+	env, err := pgEnv("postgres://llmgw:p@10.0.0.149:5432,10.0.0.150:5432/llmgw?sslmode=verify-full&sslrootcert=%2Fetc%2Fllmgw%2Fpki%2Fca.crt&target_session_attrs=read-write")
+	if err != nil {
+		t.Fatal(err)
+	}
+	got := strings.Join(env, " ")
+	for _, want := range []string{"PGSSLMODE=verify-full", "PGSSLROOTCERT=/etc/llmgw/pki/ca.crt"} {
+		if !strings.Contains(got, want) {
+			t.Errorf("%s is not in %s", want, got)
+		}
+	}
+}

internal/cli/clustersettings.go

@@ -237,8 +237,12 @@ func (w *settingsWatch) check(ctx context.Context) {
 // again: what it reads at every use, or not at all.
 var liveSettings = []string{
 	"guard",
+	// Stopping and starting the service (llmgw stop, llmgw start).
+	"serving",
 	// Sent to the nodes and this machine's firewall unit as they change.
 	"firewall", "firewall_networks",
+	// Sent to the nodes' agents as they change.
+	"agents",
 	// Read by server init, node add and backups, not by the server.
 	"backup_recipients", "backup_copy_to",
 }

internal/cli/ctl.go

@@ -116,6 +116,18 @@ func (c *ctlClient) do(ctx context.Context, method, path string, body, out any)
 	if out == nil {
 		return nil
 	}
+	// A stream, as it comes: NDJSON a long run writes line by line.
+	if w, ok := out.(io.Writer); ok {
+		_, err := io.Copy(w, resp.Body)
+		return err
+	}
+	// An older control plane answers a path it does not know with its
+	// dashboard: a page, which read as JSON said only "invalid character '<'".
+	if strings.HasPrefix(resp.Header.Get("Content-Type"), "text/html") {
+		return fmt.Errorf("%s %s answered with a page, not the API: the control plane answering "+
+			"(the leader, through a standby) does not know it — is it on an older release? "+
+			"llmgw ctl status says which it runs", method, path)
+	}
 	return json.NewDecoder(resp.Body).Decode(out)
 }
 
@@ -168,11 +180,13 @@ func newCtlCmd() *cobra.Command {
 		newJobsCmd(client, &asJSON),
 		newEngineCheckCmd(),
 		newEngineBuildCmd(client),
-		newToolCheckCmd(),
+		newToolCheckCmd(client),
 		newWeightsCmd(client, &asJSON),
 		newCapacityCmd(client, &asJSON),
 		newQuotaCmd(client, &asJSON),
 		newGuardCmd(client, &asJSON),
+		newKBCmd(client, &asJSON),
+		newTrainCmd(client, &asJSON),
 		newConfigCmd(client, &asJSON),
 		newAuditCmd(client, &asJSON),
 		newNotifyCmd(client, &asJSON),
@@ -225,8 +239,8 @@ func newNodesCmd(client func() *ctlClient, asJSON *bool) *cobra.Command {
 					// What the node can run, which is the first thing to look
 					// at when a model will not place and the hardware looks
 					// fine.
-					join(n.Engines),
-					classesColumn(n), join(n.Fabrics), n.Replicas, n.LastSeenAgo)
+					enginesColumn(n),
+					classesColumn(n), fabricsColumn(n), n.Replicas, n.LastSeenAgo)
 			}
 			_ = w.Flush()
 
@@ -443,7 +457,7 @@ func newReplicasCmd(client func() *ctlClient, asJSON *bool) *cobra.Command {
 			}
 
 			w := tab()
-			fmt.Fprintln(w, "REPLICA\tMODEL\tSTATE\tPOOL\tNODES\tAGE\tIDLE\tENDPOINT")
+			fmt.Fprintln(w, "REPLICA\tMODEL\tSTATE\tPOOL\tON\tNODES\tAGE\tIDLE\tENDPOINT")
 			for _, r := range reps {
 				var nodes []string
 				for _, s := range r.Shards {
@@ -461,8 +475,14 @@ func newReplicasCmd(client func() *ctlClient, asJSON *bool) *cobra.Command {
 				if r.Restart != "" {
 					state += " (restart " + r.Restart + ")"
 				}
-				fmt.Fprintf(w, "%s\t%s\t%s\t%s\t%s\t%s\t%s\t%s\n",
-					r.ID, r.Model, state, r.Pool, join(nodes),
+				// What the engine computes on, by its own account: the first
+				// shard's, which serves.
+				on := "-"
+				if len(r.Shards) > 0 {
+					on = r.Shards[0].Compute.Short()
+				}
+				fmt.Fprintf(w, "%s\t%s\t%s\t%s\t%s\t%s\t%s\t%s\t%s\n",
+					r.ID, r.Model, state, r.Pool, on, join(nodes),
 					(time.Duration(r.AgeSec) * time.Second).String(),
 					idleColumn(r.IdleSec), r.Endpoint)
 				// A line rather than a column, because most replicas are not
@@ -472,7 +492,7 @@ func newReplicasCmd(client func() *ctlClient, asJSON *bool) *cobra.Command {
 					if s.Fetch == nil {
 						continue
 					}
-					fmt.Fprintf(w, "  %s: fetching %s\t\t\t\t\t\t\t\n",
+					fmt.Fprintf(w, "  %s: fetching %s\t\t\t\t\t\t\t\t\n",
 						s.Node, fetchColumn(s.Fetch))
 				}
 			}
@@ -583,7 +603,42 @@ func emitJSON(v any) error {
 // more than one or MIG instances among them: a card cut into two 3g.40gb
 // instances beside a whole card is "a100-80g-mig-3g.40gb×2,a100-80g×1", where
 // the GPUS column's 3 alone would read as three of something.
+// fabricsColumn is the node's fabrics, and when they are not all up, which
+// adapters are down: what keeps it out of replicas spanning nodes.
+func fabricsColumn(n adminapi.NodeView) string {
+	s := join(n.Fabrics)
+	if n.MultiNode != "" && n.MultiNode != "ready" {
+		s += " (" + n.MultiNode + ")"
+	}
+	return s
+}
+
+// enginesColumn is each engine with its version: llamacpp b10991.
+func enginesColumn(n adminapi.NodeView) string {
+	if len(n.EngineVersions) == 0 {
+		return join(n.Engines)
+	}
+	parts := make([]string, 0, len(n.Engines))
+	for _, e := range n.Engines {
+		if v := n.EngineVersions[e]; v != "" && v != "unknown" {
+			e += " " + v
+		}
+		parts = append(parts, e)
+	}
+	return strings.Join(parts, ",")
+}
+
+// classesColumn is the node's GPU classes, with their compute capability
+// when they report one: 6000-24g (7.5).
 func classesColumn(n adminapi.NodeView) string {
+	s := classesOnly(n)
+	if len(n.ComputeCapabilities) > 0 && s != "-" {
+		s += " (" + strings.Join(n.ComputeCapabilities, ",") + ")"
+	}
+	return s
+}
+
+func classesOnly(n adminapi.NodeView) string {
 	counted := len(n.ClassGPUs) > 1
 	for _, c := range n.ClassGPUs {
 		counted = counted || c.MIG

internal/cli/ctl_config_test.go

@@ -0,0 +1,22 @@
+package cli
+
+import (
+	"net/http"
+	"net/http/httptest"
+	"strings"
+	"testing"
+)
+
+func TestAPageInsteadOfTheAPIIsSaidPlainly(t *testing.T) {
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
+		w.Header().Set("Content-Type", "text/html; charset=utf-8")
+		_, _ = w.Write([]byte("<!doctype html><title>llmgw</title>"))
+	}))
+	defer srv.Close()
+	c := newCtlClient(strings.TrimPrefix(srv.URL, "http://"), "")
+	var out map[string]any
+	err := c.get(t.Context(), "/api/v1/settings", &out)
+	if err == nil || !strings.Contains(err.Error(), "older release") || strings.Contains(err.Error(), "invalid character") {
+		t.Errorf("%v", err)
+	}
+}

internal/cli/ctl_request.go

@@ -76,8 +76,11 @@ func newRequestCmd(client func() *ctlClient, asJSON *bool) *cobra.Command {
 				fmt.Fprintf(out, "tags      %s\n", strings.Join(kv, ", "))
 			}
 			if v.GuardAction != "" {
-				fmt.Fprintf(out, "guard     %s by policy %s (%s), %d ms\n", v.GuardAction, v.GuardPolicy,
-					or(v.GuardCategory, "-"), v.GuardMs)
+				what := or(v.GuardCategory, "-")
+				if v.GuardRule != "" {
+					what += ": " + v.GuardRule
+				}
+				fmt.Fprintf(out, "guard     %s by policy %s (%s), %d ms\n", v.GuardAction, v.GuardPolicy, what, v.GuardMs)
 			} else if v.GuardMs > 0 {
 				fmt.Fprintf(out, "guard     allowed, %d ms\n", v.GuardMs)
 			}

internal/cli/ctl_restart.go

@@ -12,9 +12,10 @@ import (
 
 func newRestartCmd(client func() *ctlClient) *cobra.Command {
 	var (
-		models  []string
-		all     bool
-		inPlace bool
+		models             []string
+		all                bool
+		inPlace            bool
+		surge, unavailable int
 	)
 	c := &cobra.Command{
 		Use:   "restart [REPLICA...]",
@@ -31,6 +32,10 @@ func newRestartCmd(client func() *ctlClient) *cobra.Command {
 			"profile edited to need more than any node has — the restart is given up\n" +
 			"and the old replica keeps serving: llmgw ctl replicas says so, and why.\n\n" +
 			"--in-place stops each replica first, whatever room there is.\n\n" +
+			"A model's replicas are replaced one at a time. --max-surge N replaces N at\n" +
+			"once, each beside itself, needing room for N more; --max-unavailable N\n" +
+			"stops N at once first, needing no room and leaving N fewer serving while\n" +
+			"they load. Together, the first N are stopped first and the rest surged.\n\n" +
 			"  llmgw ctl restart qwen2.5-7b-3f2a91c0\n" +
 			"  llmgw ctl restart --model qwen2.5-7b\n" +
 			"  llmgw ctl restart --all",
@@ -38,7 +43,8 @@ func newRestartCmd(client func() *ctlClient) *cobra.Command {
 			if len(args) == 0 && len(models) == 0 && !all {
 				return errors.New("name replicas, --model, or --all")
 			}
-			body, _ := json.Marshal(adminapi.RestartRequest{Replicas: args, Models: models, All: all, InPlace: inPlace})
+			body, _ := json.Marshal(adminapi.RestartRequest{Replicas: args, Models: models, All: all, InPlace: inPlace,
+				MaxSurge: surge, MaxUnavailable: unavailable})
 			var res adminapi.RestartResult
 			if err := client().post(cmd.Context(), "/api/v1/restart", body, &res); err != nil {
 				return err
@@ -50,6 +56,9 @@ func newRestartCmd(client func() *ctlClient) *cobra.Command {
 	c.Flags().StringArrayVar(&models, "model", nil, "every replica of this model; repeatable")
 	c.Flags().BoolVar(&all, "all", false, "every ready replica in the cluster")
 	c.Flags().BoolVar(&inPlace, "in-place", false, "stop each replica before starting its replacement")
+	c.Flags().IntVar(&surge, "max-surge", 0, "replicas of a model replaced beside themselves at once (default 1)")
+	c.Flags().IntVar(&unavailable, "max-unavailable", 0,
+		"replicas of a model stopped first at once, each out while its replacement loads")
 	return c
 }
 

internal/cli/doctor.go

@@ -63,9 +63,10 @@ that file when the report is not enough to go on.`,
 				enc := json.NewEncoder(os.Stdout)
 				enc.SetIndent("", "  ")
 				return enc.Encode(struct {
-					Findings []doctor.Finding `json:"findings"`
-					Facts    doctor.Facts     `json:"facts"`
-				}{findings, facts})
+					Findings  []doctor.Finding `json:"findings"`
+					Readiness doctor.Readiness `json:"readiness"`
+					Facts     doctor.Facts     `json:"facts"`
+				}{findings, doctor.ReadinessOf(findings), facts})
 			}
 			printDoctor(os.Stdout, facts, findings, verbose)
 			for _, f := range findings {
@@ -101,12 +102,26 @@ func printDoctor(out io.Writer, f doctor.Facts, findings []doctor.Finding, verbo
 	fmt.Fprintf(out, "%s · %s %s %s · driver %s\n\n", f.Hostname, f.Host.OSID, f.Host.OSVersion,
 		f.Host.Arch, orDash(f.Host.NVIDIADriver))
 
-	for _, x := range findings {
-		fmt.Fprintf(out, "%-4s  %-6s  %s\n", x.Level, x.Area, x.Message)
-		if x.Fix != "" && (x.Level == doctor.Fail || x.Level == doctor.Warn) {
-			fmt.Fprintf(out, "              → %s\n", x.Fix)
+	// The node first; then what replicas spanning nodes would need, which
+	// keeps a node out only when a fabric was declared for it.
+	multi := false
+	for _, pass := range []bool{false, true} {
+		for _, x := range findings {
+			if x.MultiNode() != pass {
+				continue
+			}
+			if pass && !multi {
+				fmt.Fprintln(out, "\nMulti-node (fabrics, RDMA) — for replicas spanning nodes only:")
+				multi = true
+			}
+			fmt.Fprintf(out, "%-4s  %-6s  %s\n", x.Level, x.Area, x.Message)
+			if x.Fix != "" && (x.Level == doctor.Fail || x.Level == doctor.Warn) {
+				fmt.Fprintf(out, "              → %s\n", x.Fix)
+			}
 		}
 	}
+	r := doctor.ReadinessOf(findings)
+	fmt.Fprintf(out, "\nThis node:  %s\nMulti-node: %s\n", r.Node, r.MultiNode)
 
 	w := tabwriter.NewWriter(out, 0, 0, 2, ' ', 0)
 	if len(f.GPUs) > 0 {

internal/cli/enginecheck.go

@@ -4,10 +4,15 @@ import (
 	"context"
 	"fmt"
 	"io"
+	"net"
+	"net/http"
+	"os"
 	"os/exec"
 	"path/filepath"
 	"sort"
+	"strconv"
 	"strings"
+	"sync"
 	"time"
 
 	"github.com/spf13/cobra"
@@ -22,6 +27,7 @@ func newEngineCheckCmd() *cobra.Command {
 		nodes   int
 		tp      int
 		perNode int
+		model   string
 	)
 
 	cmd := &cobra.Command{
@@ -31,6 +37,10 @@ func newEngineCheckCmd() *cobra.Command {
 			"engines rename and remove flags between versions. When that happens the\n" +
 			"failure arrives minutes into a model load, in an engine's own error\n" +
 			"message, on a node nobody was watching.\n\n" +
+			"For llama.cpp on a machine with NVIDIA GPUs it then loads a small GGUF\n" +
+			"(--model, or the smallest in the weights cache) with every layer on the GPU\n" +
+			"and asks for one token: a build whose CUDA backend does not load computes on\n" +
+			"the CPU, and that is where it is caught.\n\n" +
 			"This asks the engine itself. It runs the engine's --help and checks every\n" +
 			"flag llmgw would pass against it, which needs no GPU — a CLI prints its\n" +
 			"help on any machine. Run it after upgrading an engine.\n\n" +
@@ -49,6 +59,11 @@ func newEngineCheckCmd() *cobra.Command {
 			// rarely on PATH: a check that cannot find the engine reports
 			// nothing rather than reporting a problem.
 			cfg := engines.Config{}
+			// The build the agent installed, when none is named: the newest
+			// under /opt/llmgw/engines, which is where every node keeps it.
+			if bin == "" && name == "llamacpp" {
+				bin = installedLlama()
+			}
 			if bin != "" {
 				switch name {
 				case "llamacpp":
@@ -135,6 +150,11 @@ func newEngineCheckCmd() *cobra.Command {
 			var deviceErr error
 			if fb, ok := eng.(engines.CPUFallback); ok && help == "" {
 				deviceErr = checkDevices(cmd.Context(), name, fb, out)
+				// Seeing a GPU is not computing on it: a small model loaded
+				// on it and asked for a token is.
+				if b, ok := eng.(interface{ Binary() string }); ok && deviceErr == nil && model != "none" {
+					deviceErr = computeOnGPU(cmd.Context(), b.Binary(), model, out)
+				}
 			}
 
 			fmt.Fprintf(out, "%s: %d flags checked", name, checked)
@@ -160,7 +180,9 @@ func newEngineCheckCmd() *cobra.Command {
 
 	f := cmd.Flags()
 	f.StringVar(&help, "help-output", "", "a saved --help to check against, instead of running the engine")
-	f.StringVar(&bin, "bin", "", "where the engine lives, when it is not on PATH")
+	f.StringVar(&bin, "bin", "", "where the engine lives (default: the build the agent installed, else PATH)")
+	f.StringVar(&model, "model", "", "a GGUF file to load on the GPU and ask for a token (default: the smallest "+
+		"in this node's weights cache; none skips it)")
 	f.IntVar(&nodes, "nodes", 2, "shape the command for a replica of this many nodes")
 	f.IntVar(&tp, "tp", 8, "tensor parallel degree")
 	f.IntVar(&perNode, "gpus-per-node", 4, "GPUs each node contributes")
@@ -286,3 +308,146 @@ func checkDevices(ctx context.Context, name string, fb engines.CPUFallback, out
 	fmt.Fprintf(out, "computes on: %s\n", strings.Join(devs, "; "))
 	return nil
 }
+
+// llamaRoot is where the agent installs llama.cpp builds.
+var llamaRoot = "/opt/llmgw/engines/llamacpp"
+
+// installedLlama is the llama-server the agent installed here: under
+// llamaRoot/<version-variant>/llama-<build>/, or one level up for older
+// layouts; on a machine with NVIDIA GPUs a CUDA build first, then the newest.
+func installedLlama() string {
+	var found []string
+	for _, pat := range []string{"*/llama-server", "*/*/llama-server"} {
+		m, _ := filepath.Glob(filepath.Join(llamaRoot, pat))
+		found = append(found, m...)
+	}
+	if len(found) == 0 {
+		return ""
+	}
+	sort.Strings(found)
+	if _, err := exec.LookPath("nvidia-smi"); err == nil {
+		for i := len(found) - 1; i >= 0; i-- {
+			if strings.Contains(found[i], "cuda") {
+				return found[i]
+			}
+		}
+	}
+	return found[len(found)-1]
+}
+
+// weightDirs are where a node keeps models it has fetched.
+var weightDirs = []string{"/var/lib/llmgw-weights", "/var/lib/llmgw-models", "/srv/models"}
+
+// smallestGGUF is the smallest GGUF file under dirs, for a check that only
+// needs something to load.
+func smallestGGUF(dirs []string) string {
+	best, size := "", int64(0)
+	for _, d := range dirs {
+		_ = filepath.WalkDir(d, func(p string, e os.DirEntry, err error) error {
+			if err != nil || e.IsDir() || !strings.HasSuffix(strings.ToLower(p), ".gguf") ||
+				strings.Contains(strings.ToLower(filepath.Base(p)), "mmproj") {
+				return nil
+			}
+			if info, err := e.Info(); err == nil && (best == "" || info.Size() < size) {
+				best, size = p, info.Size()
+			}
+			return nil
+		})
+	}
+	return best
+}
+
+// computeOnGPU loads a model with every layer on the GPU, asks for one
+// token, and reads from llama.cpp's own output where the layers went.
+func computeOnGPU(ctx context.Context, bin, model string, out io.Writer) error {
+	if _, err := exec.LookPath("nvidia-smi"); err != nil {
+		return nil
+	}
+	if model == "" {
+		if model = smallestGGUF(weightDirs); model == "" {
+			fmt.Fprintln(out, "no GGUF in this node's weights cache to compute with; --model PATH runs one")
+			return nil
+		}
+	}
+	ln, err := net.Listen("tcp", "127.0.0.1:0")
+	if err != nil {
+		return err
+	}
+	port := ln.Addr().(*net.TCPAddr).Port
+	_ = ln.Close()
+	ctx, cancel := context.WithTimeout(ctx, 3*time.Minute)
+	defer cancel()
+	var logs syncBuffer
+	c := exec.CommandContext(ctx, bin, "-m", model, "--host", "127.0.0.1", "--port", strconv.Itoa(port), //nolint:gosec // the engine's own binary
+		"--n-gpu-layers", "999", "--ctx-size", "512", "--parallel", "1",
+		// Where the layers went is said only at this verbosity.
+		"-lv", "4")
+	c.Stdout, c.Stderr = &logs, &logs
+	if err := c.Start(); err != nil {
+		return fmt.Errorf("starting %s: %w", bin, err)
+	}
+	defer func() { _ = c.Process.Kill(); _ = c.Wait() }()
+	base := "http://127.0.0.1:" + strconv.Itoa(port)
+	start := time.Now()
+	for {
+		resp, err := http.Get(base + "/health") //nolint:gosec,noctx // loopback, bounded by the loop
+		if err == nil {
+			_ = resp.Body.Close()
+			if resp.StatusCode == http.StatusOK {
+				break
+			}
+		}
+		if ctx.Err() != nil {
+			return fmt.Errorf("%s did not come up with %s: %s", bin, filepath.Base(model), tailLines(logs.String(), 5))
+		}
+		time.Sleep(500 * time.Millisecond)
+	}
+	req, _ := http.NewRequestWithContext(ctx, http.MethodPost, base+"/completion",
+		strings.NewReader(`{"prompt":"Hello","n_predict":1}`))
+	req.Header.Set("Content-Type", "application/json")
+	resp, err := http.DefaultClient.Do(req)
+	if err != nil {
+		return fmt.Errorf("asking it for a token: %w", err)
+	}
+	_ = resp.Body.Close()
+	took := time.Since(start)
+	comp, said := engines.ReadLlamaCompute(strings.Split(logs.String(), "\n"))
+	if !said {
+		fmt.Fprintf(out, "computed a token with %s, but llama.cpp did not say where\n", filepath.Base(model))
+		return nil
+	}
+	if comp.OnTheCPU() {
+		return fmt.Errorf("%s computed on the CPU though every layer was asked for on the GPU (%d of %d offloaded): "+
+			"its CUDA backend did not load; ldd on its libggml-cuda.so names the missing library",
+			filepath.Base(model), comp.Offloaded, comp.Total)
+	}
+	fmt.Fprintf(out, "computed a token on %s: %s, %d of %d layers offloaded, in %s\n", comp.Backend,
+		strings.Join(comp.Devices, "; "), comp.Offloaded, comp.Total, took.Round(100*time.Millisecond))
+	return nil
+}
+
+func tailLines(s string, n int) string {
+	l := strings.Split(strings.TrimSpace(s), "\n")
+	if len(l) > n {
+		l = l[len(l)-n:]
+	}
+	return strings.Join(l, " | ")
+}
+
+// syncBuffer is a buffer two writers share.
+type syncBuffer struct {
+	mu sync.Mutex
+	b  strings.Builder
+}
+
+func (s *syncBuffer) Write(p []byte) (int, error) {
+	s.mu.Lock()
+	defer s.mu.Unlock()
+	return s.b.Write(p)
+}
+
+func (s *syncBuffer) String() string {
+	s.mu.Lock()
+	defer s.mu.Unlock()
+	return s.b.String()
+}

internal/cli/enginecheck_gpu_test.go

@@ -0,0 +1,86 @@
+package cli
+
+import (
+	"bytes"
+	"fmt"
+	"net/http"
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+)
+
+// TestMain lets the test binary stand in for llama-server: run with
+// LLMGW_FAKE_LLAMA set, it prints what llama.cpp prints and answers.
+func TestMain(m *testing.M) {
+	if mode := os.Getenv("LLMGW_FAKE_LLAMA"); mode != "" {
+		fakeLlama(mode)
+		return
+	}
+	os.Exit(m.Run())
+}
+
+func fakeLlama(mode string) {
+	port := ""
+	for i, a := range os.Args {
+		if a == "--port" && i+1 < len(os.Args) {
+			port = os.Args[i+1]
+		}
+	}
+	if mode == "gpu" {
+		fmt.Println("llama_model_load_from_file_impl: using device CUDA0 (Quadro RTX 6000) - 23800 MiB free")
+		fmt.Println("load_tensors: offloaded 25/25 layers to GPU")
+	} else {
+		fmt.Println("warning: no usable GPU found, --gpu-layers option will be ignored")
+		fmt.Println("load_tensors: offloaded 0/25 layers to GPU")
+	}
+	http.HandleFunc("/health", func(w http.ResponseWriter, _ *http.Request) { w.WriteHeader(http.StatusOK) })
+	http.HandleFunc("/completion", func(w http.ResponseWriter, _ *http.Request) { _, _ = w.Write([]byte(`{"content":"!"}`)) })
+	_ = http.ListenAndServe("127.0.0.1:"+port, nil) //nolint:gosec // a test's stand-in
+}
+
+func TestEngineCheckComputesATokenOnTheGPU(t *testing.T) {
+	// nvidia-smi present, so the check runs.
+	dir := t.TempDir()
+	smi := filepath.Join(dir, "nvidia-smi")
+	if err := os.WriteFile(smi, []byte("#!/bin/sh\nexit 0\n"), 0o755); err != nil {
+		t.Fatal(err)
+	}
+	t.Setenv("PATH", dir+string(os.PathListSeparator)+os.Getenv("PATH"))
+	model := filepath.Join(dir, "tiny.gguf")
+	_ = os.WriteFile(model, []byte("GGUF"), 0o644)
+	self, _ := os.Executable()
+
+	t.Setenv("LLMGW_FAKE_LLAMA", "gpu")
+	var out bytes.Buffer
+	if err := computeOnGPU(t.Context(), self, model, &out); err != nil || !strings.Contains(out.String(), "on CUDA") ||
+		!strings.Contains(out.String(), "25 of 25") {
+		t.Fatalf("on the GPU: %v %s", err, out.String())
+	}
+
+	t.Setenv("LLMGW_FAKE_LLAMA", "cpu")
+	out.Reset()
+	if err := computeOnGPU(t.Context(), self, model, &out); err == nil || !strings.Contains(err.Error(), "computed on the CPU") {
+		t.Fatalf("a build that falls back to the CPU: %v %s", err, out.String())
+	}
+}
+
+func TestEngineCheckFindsTheBuildTheAgentInstalled(t *testing.T) {
+	old := llamaRoot
+	llamaRoot = t.TempDir()
+	t.Cleanup(func() { llamaRoot = old })
+	if got := installedLlama(); got != "" {
+		t.Fatalf("nothing installed, found %q", got)
+	}
+	for _, p := range []string{"b10991-cpu/llama-b10991/llama-server", "b10991-cuda-13.3/llama-b10991/llama-server"} {
+		full := filepath.Join(llamaRoot, p)
+		_ = os.MkdirAll(filepath.Dir(full), 0o755)
+		_ = os.WriteFile(full, []byte("x"), 0o755)
+	}
+	dir := t.TempDir()
+	_ = os.WriteFile(filepath.Join(dir, "nvidia-smi"), []byte("#!/bin/sh\n"), 0o755)
+	t.Setenv("PATH", dir)
+	if got := installedLlama(); !strings.Contains(got, "cuda-13.3/llama-b10991/llama-server") {
+		t.Errorf("with GPUs, not the CUDA build: %q", got)
+	}
+}

internal/cli/firewallcmd.go

@@ -7,12 +7,14 @@ import (
 	"fmt"
 	"net/netip"
 	"os"
+	"strings"
 	"time"
 
 	"github.com/spf13/cobra"
 
 	"gitea.mynodes.xyz/ilker/llmgw/internal/config"
 	"gitea.mynodes.xyz/ilker/llmgw/internal/firewall"
+	"gitea.mynodes.xyz/ilker/llmgw/internal/mirror"
 )
 
 // firewallStateFile is where a control plane writes what the cluster's
@@ -20,11 +22,9 @@ import (
 // the server itself runs as llmgw and cannot change the ruleset.
 const firewallStateFile = "/var/lib/llmgw/firewall.json"
 
-// firewallState is what is written there.
-type firewallState struct {
-	Enabled  bool     `json:"enabled"`
-	Networks []string `json:"networks,omitempty"`
-}
+// firewallState is what is written there: the same as the mirror is told
+// and keeps beside its models.
+type firewallState = mirror.FirewallState
 
 // writeFirewallState records the settings' firewall for this machine's sync
 // unit. Without the state directory — a server not run by systemd — there
@@ -46,25 +46,34 @@ func writeFirewallState(cfg config.Server) error {
 
 // The units that apply it: a path unit watching the file, and the oneshot it
 // starts.
-const (
-	firewallSyncPathUnit = "llmgw-firewall-sync.path"
-	firewallSyncUnitFile = `[Unit]
-Description=llmgw firewall: apply the cluster's settings on this control plane
+const firewallSyncPathUnit = "llmgw-firewall-sync.path"
+
+// firewallSyncUnits are the two units for a state file: on a control plane
+// the one its server writes, on a mirror the one the control plane has the
+// mirror write.
+func firewallSyncUnits(state string) (service, pathUnit string) {
+	arg := ""
+	if state != firewallStateFile {
+		arg = " --state " + state
+	}
+	service = `[Unit]
+Description=llmgw firewall: apply the cluster's settings on this machine
 
 [Service]
 Type=oneshot
-ExecStart=/usr/local/bin/llmgw firewall sync
+ExecStart=/usr/local/bin/llmgw firewall sync` + arg + `
 `
-	firewallSyncPathFile = `[Unit]
-Description=llmgw firewall: follow the cluster's settings on this control plane
+	pathUnit = `[Unit]
+Description=llmgw firewall: follow the cluster's settings on this machine
 
 [Path]
-PathChanged=` + firewallStateFile + `
+PathChanged=` + state + `
 
 [Install]
 WantedBy=multi-user.target
 `
-)
+	return service, pathUnit
+}
 
 func newFirewallCmd() *cobra.Command {
 	c := &cobra.Command{
@@ -76,6 +85,7 @@ func newFirewallCmd() *cobra.Command {
 			"which " + firewallSyncPathUnit + " applies with 'llmgw firewall sync'.",
 	}
 	var install bool
+	state := firewallStateFile
 	sync := &cobra.Command{
 		Use:   "sync",
 		Short: "Apply what the cluster's settings say of the firewall, here (root)",
@@ -85,21 +95,21 @@ func newFirewallCmd() *cobra.Command {
 				return errors.New("the firewall is changed as root: run it with sudo")
 			}
 			if install {
-				if err := installFirewallSync(cmd.Context(), firewall.Local{}); err != nil {
+				if err := installFirewallSync(cmd.Context(), firewall.Local{}, state); err != nil {
 					return err
 				}
 				fmt.Fprintln(cmd.OutOrStdout(), firewallSyncPathUnit+" follows the cluster's settings here")
-				if _, err := os.Stat(firewallStateFile); err != nil {
-					return nil // the server writes it when it starts
+				if _, err := os.Stat(state); err != nil {
+					return nil // written when the settings are first told
 				}
 			}
-			b, err := os.ReadFile(firewallStateFile)
+			b, err := os.ReadFile(state)
 			if err != nil {
 				return err
 			}
 			var st firewallState
 			if err := json.Unmarshal(b, &st); err != nil {
-				return fmt.Errorf("%s: %w", firewallStateFile, err)
+				return fmt.Errorf("%s: %w", state, err)
 			}
 			ctx, cancel := context.WithTimeout(cmd.Context(), 2*time.Minute)
 			defer cancel()
@@ -112,14 +122,17 @@ func newFirewallCmd() *cobra.Command {
 		},
 	}
 	sync.Flags().BoolVar(&install, "install", false, "first install the units that run this whenever the settings change")
+	sync.Flags().StringVar(&state, "state", firewallStateFile,
+		"the file the settings are written to: a control plane's, or "+mirror.FirewallFile+" in a mirror's directory")
 	c.AddCommand(sync)
 	return c
 }
 
 // installFirewallSync writes the sync units on a machine and has the path
 // unit follow the state file.
-func installFirewallSync(ctx context.Context, r firewall.Runner) error {
-	body := firewallSyncUnitFile + "\x00" + firewallSyncPathFile
+func installFirewallSync(ctx context.Context, r firewall.Runner, state string) error {
+	service, pathUnit := firewallSyncUnits(state)
+	body := service + "\x00" + pathUnit
 	out, err := r.Run(ctx, `set -e
 t=$(mktemp -d); trap 'rm -rf "$t"' EXIT
 cat > "$t/all"
@@ -172,5 +185,20 @@ func syncFirewall(ctx context.Context, r firewall.Runner, st firewallState) (str
 			return "", err
 		}
 	}
+	// The cluster's machines: a node on a network this one has not let in
+	// yet, which the mirror's files are for.
+	if len(st.Toward) > 0 {
+		added, err := firewall.Admit(ctx, r, st.Toward)
+		if err != nil {
+			return "", err
+		}
+		if len(added) > 0 {
+			shown := make([]string, len(added))
+			for i, p := range added {
+				shown[i] = p.String()
+			}
+			return "llmgw's firewall is on here, and now lets in " + strings.Join(shown, ", "), nil
+		}
+	}
 	return "llmgw's firewall is on here", nil
 }

internal/cli/firewallcmd_test.go

@@ -2,8 +2,11 @@ package cli
 
 import (
 	"context"
+	"net/netip"
 	"strings"
 	"testing"
+
+	"gitea.mynodes.xyz/ilker/llmgw/internal/firewall"
 )
 
 type scriptLog struct {
@@ -49,3 +52,32 @@ func TestAControlPlaneFollowsTheFirewallSetting(t *testing.T) {
 		t.Errorf("on: %q %v", msg, err)
 	}
 }
+
+func TestAMachineLetsInTheClusterMachinesItIsTold(t *testing.T) {
+	// The rules hold 10.0.0.0/24; a node on 192.168.7.0/24 is let in, as
+	// the route to it says, and one already in is not asked about.
+	r := &scriptLog{reply: func(s string) string {
+		switch {
+		case strings.Contains(s, "echo loaded"):
+			return "loaded\n"
+		case strings.Contains(s, "cat /etc/llmgw/firewall.nft"):
+			return firewallRulesFor("10.0.0.0/24")
+		case strings.Contains(s, "ip -o route get 192.168.7.20"):
+			return "192.168.7.20 dev eth1 src 192.168.7.1 uid 0\n2: eth1    inet 192.168.7.1/24 brd 192.168.7.255 scope global eth1\n"
+		case strings.Contains(s, "nft -f"):
+			return "changed\n"
+		}
+		return ""
+	}}
+	msg, err := syncFirewall(context.Background(), r, firewallState{Enabled: true, Toward: []string{"10.0.0.147", "192.168.7.20"}})
+	if err != nil || !strings.Contains(msg, "192.168.7.0/24") {
+		t.Fatalf("%q %v", msg, err)
+	}
+	if strings.Contains(strings.Join(r.scripts, "\n"), "route get 10.0.0.147") {
+		t.Error("asked the route to a machine already let in")
+	}
+}
+
+func firewallRulesFor(nets string) string {
+	return firewall.Script(firewall.Spec{Ports: []firewall.Port{firewall.MirrorFiles}, Allow: []netip.Prefix{netip.MustParsePrefix(nets)}})
+}

internal/cli/firewallpub.go

@@ -0,0 +1,146 @@
+package cli
+
+import (
+	"context"
+	"encoding/json"
+	"log/slog"
+	"net"
+	"net/netip"
+	"net/url"
+	"slices"
+	"sync"
+	"time"
+
+	"gitea.mynodes.xyz/ilker/llmgw/internal/config"
+	"gitea.mynodes.xyz/ilker/llmgw/internal/mirror"
+	"gitea.mynodes.xyz/ilker/llmgw/internal/store"
+)
+
+// firewallStateKey is where, in the store, the firewall's state is kept
+// for the store's members to read.
+const firewallStateKey = "firewall"
+
+// firewallPublisher tells the machines that cannot ask a control plane
+// what the cluster's settings say of their firewall: the mirror, over its
+// API, and the store's members, through the store they hold a copy of.
+// Nodes are told over their agents' streams and control planes by their
+// own servers. On the leader only, at start, at every change to the
+// firewall's settings, and every few minutes for nodes come and gone.
+type firewallPublisher struct {
+	settings func() config.Server
+	st       *store.DB
+	nodes    func() []string
+	leader   func() bool
+	log      *slog.Logger
+
+	mu     sync.Mutex
+	mirror *mirror.Client
+	kick   chan struct{}
+}
+
+func newFirewallPublisher(settings func() config.Server, st *store.DB, nodes func() []string,
+	leader func() bool, log *slog.Logger) *firewallPublisher {
+	return &firewallPublisher{settings: settings, st: st, nodes: nodes, leader: leader, log: log,
+		kick: make(chan struct{}, 1)}
+}
+
+// setMirror has the mirror told too.
+func (p *firewallPublisher) setMirror(m *mirror.Client) {
+	p.mu.Lock()
+	p.mirror = m
+	p.mu.Unlock()
+	p.Kick()
+}
+
+// Kick has the state worked out and told again now.
+func (p *firewallPublisher) Kick() {
+	select {
+	case p.kick <- struct{}{}:
+	default:
+	}
+}
+
+func (p *firewallPublisher) run(ctx context.Context, every time.Duration) {
+	t := time.NewTicker(every)
+	defer t.Stop()
+	for {
+		p.once(ctx)
+		select {
+		case <-ctx.Done():
+			return
+		case <-t.C:
+		case <-p.kick:
+		}
+	}
+}
+
+// state is what the settings say now, with the cluster's machines.
+func (p *firewallPublisher) state(ctx context.Context) mirror.FirewallState {
+	cfg := p.settings()
+	st := mirror.FirewallState{Enabled: cfg.Firewall.Enabled, Networks: slices.Clone(cfg.FirewallNetworks)}
+	var toward []string
+	if p.nodes != nil {
+		toward = append(toward, p.nodes()...)
+	}
+	if p.st != nil {
+		if cps, err := p.st.ControlPlanes(ctx); err == nil {
+			for _, c := range cps {
+				toward = append(toward, c.URL)
+			}
+		}
+	}
+	for _, t := range toward {
+		if a, ok := hostAddr(t); ok {
+			st.Toward = append(st.Toward, a)
+		}
+	}
+	slices.Sort(st.Toward)
+	st.Toward = slices.Compact(st.Toward)
+	return st
+}
+
+// hostAddr is the address in an address, host:port or URL; names are not
+// looked up here.
+func hostAddr(s string) (string, bool) {
+	if u, err := url.Parse(s); err == nil && u.Host != "" {
+		s = u.Hostname()
+	} else if h, _, err := net.SplitHostPort(s); err == nil {
+		s = h
+	}
+	a, err := netip.ParseAddr(s)
+	if err != nil {
+		return "", false
+	}
+	return a.Unmap().String(), true
+}
+
+func (p *firewallPublisher) once(ctx context.Context) {
+	if p.leader != nil && !p.leader() {
+		return
+	}
+	ctx, cancel := context.WithTimeout(ctx, 30*time.Second)
+	defer cancel()
+	st := p.state(ctx)
+	p.mu.Lock()
+	m := p.mirror
+	p.mu.Unlock()
+	if m != nil {
+		changed, err := m.SetFirewall(ctx, st)
+		switch {
+		case err != nil:
+			p.log.Warn("could not tell the mirror what the settings say of its firewall", "err", err)
+		case changed:
+			p.log.Info("the mirror's firewall follows the cluster's settings", "enabled", st.Enabled)
+		}
+	}
+	if p.st != nil {
+		b, _ := json.Marshal(st)
+		changed, err := p.st.PutClusterState(ctx, firewallStateKey, string(b))
+		switch {
+		case err != nil:
+			p.log.Warn("could not record the firewall's state for the store's members", "err", err)
+		case changed:
+			p.log.Info("the store's members' firewall follows the cluster's settings", "enabled", st.Enabled)
+		}
+	}
+}

internal/cli/firewallpub_test.go

@@ -0,0 +1,59 @@
+package cli
+
+import (
+	"context"
+	"encoding/json"
+	"log/slog"
+	"net/http/httptest"
+	"os"
+	"path/filepath"
+	"testing"
+
+	"gitea.mynodes.xyz/ilker/llmgw/internal/config"
+	"gitea.mynodes.xyz/ilker/llmgw/internal/mirror"
+	"gitea.mynodes.xyz/ilker/llmgw/internal/store"
+	"gitea.mynodes.xyz/ilker/llmgw/internal/tlsconf"
+)
+
+func TestTheLeaderTellsTheMirrorAndTheStoreWhatTheFirewallSettingsSay(t *testing.T) {
+	ctx := context.Background()
+	st, err := store.OpenSQLite(ctx, filepath.Join(t.TempDir(), "s.db"))
+	if err != nil {
+		t.Fatal(err)
+	}
+	defer st.Close()
+	if err := st.PutControlPlane(ctx, store.ControlPlane{URL: "https://10.0.0.150:7443", Name: "ctl02"}); err != nil {
+		t.Fatal(err)
+	}
+	root := t.TempDir()
+	srv := httptest.NewServer(mirror.NewAPI(mirror.NewManager(root, slog.New(slog.DiscardHandler)), "http://m:8081").Handler())
+	defer srv.Close()
+	mc, _ := mirror.NewClient(srv.URL, tlsconf.Config{}, 0)
+
+	cfg := config.Server{Firewall: config.FirewallSettings{Enabled: false}, FirewallNetworks: []string{"192.168.5.0/24"}}
+	leader := false
+	p := newFirewallPublisher(func() config.Server { return cfg }, st,
+		func() []string { return []string{"10.0.0.147", "", "10.0.0.147", "node-by-name"} },
+		func() bool { return leader }, slog.New(slog.DiscardHandler))
+	p.setMirror(mc)
+
+	p.once(ctx)
+	if _, err := st.ClusterState(ctx, firewallStateKey); err == nil {
+		t.Error("a standby told the store")
+	}
+	leader = true
+	p.once(ctx)
+	var got mirror.FirewallState
+	v, err := st.ClusterState(ctx, firewallStateKey)
+	if err != nil || json.Unmarshal([]byte(v), &got) != nil {
+		t.Fatalf("the store holds %q %v", v, err)
+	}
+	if got.Enabled || got.Networks[0] != "192.168.5.0/24" || len(got.Toward) != 2 ||
+		got.Toward[0] != "10.0.0.147" || got.Toward[1] != "10.0.0.150" {
+		t.Errorf("state %+v", got)
+	}
+	b, err := os.ReadFile(filepath.Join(root, mirror.FirewallFile))
+	if err != nil || string(b) != v {
+		t.Errorf("the mirror holds %s, the store %s", b, v)
+	}
+}

internal/cli/guard.go

@@ -3,11 +3,13 @@ package cli
 import (
 	"encoding/json"
 	"fmt"
+	"strconv"
 	"strings"
 
 	"github.com/spf13/cobra"
 
 	"gitea.mynodes.xyz/ilker/llmgw/internal/control/adminapi"
+	"gitea.mynodes.xyz/ilker/llmgw/internal/settings"
 )
 
 func newGuardCmd(client func() *ctlClient, asJSON *bool) *cobra.Command {
@@ -24,6 +26,35 @@ func newGuardCmd(client func() *ctlClient, asJSON *bool) *cobra.Command {
 			"Targets are written org:Acme, team:Acme/ml, project:Acme/ml/training (or\n" +
 			"with IDs), key:KEY-ID or user:USER-ID. An attachment applies at once.",
 	}
+	c.AddCommand(newGuardJudgeCmd(client, asJSON))
+	for _, sw := range []struct {
+		use, short string
+		on         bool
+	}{
+		{"on", "Turn the guardrails on: every policy checks again", true},
+		{"off", "Turn the guardrails off: no request is checked, whatever the policies say", false},
+	} {
+		var note string
+		cmd := &cobra.Command{
+			Use:   sw.use,
+			Short: sw.short,
+			Long: "guard.enabled in the cluster's settings: in force on every control plane\n" +
+				"within seconds, without a restart. The policies and where they are attached\n" +
+				"stay as they are, and come back as they were when turned on again. While\n" +
+				"off, llmgw ctl status says so.",
+			Args: cobra.NoArgs,
+			RunE: func(cmd *cobra.Command, _ []string) error {
+				if note == "" {
+					note = "guardrails " + sw.use
+				}
+				return changeSettings(cmd, client, note, func(doc *settings.Doc) error {
+					return doc.SetYAML([]string{"guard", "enabled"}, strconv.FormatBool(sw.on))
+				})
+			},
+		}
+		cmd.Flags().StringVar(&note, "note", "", "why, kept with the settings' version")
+		c.AddCommand(cmd)
+	}
 	c.AddCommand(&cobra.Command{
 		Use:   "list",
 		Short: "The policies, and where each is attached",
@@ -37,6 +68,9 @@ func newGuardCmd(client func() *ctlClient, asJSON *bool) *cobra.Command {
 				return json.NewEncoder(cmd.OutOrStdout()).Encode(v)
 			}
 			out := cmd.OutOrStdout()
+			if !v.Enabled {
+				fmt.Fprintln(out, "The guardrails are OFF: no request is checked (llmgw ctl guard on).")
+			}
 			fmt.Fprintf(out, "Policies: %s\n", strings.Join(v.Policies, ", "))
 			def := strings.Join(v.Default, ", ")
 			if def == "" {

internal/cli/guard_judge.go

@@ -0,0 +1,385 @@
+package cli
+
+import (
+	"encoding/json"
+	"errors"
+	"fmt"
+	"io"
+	"slices"
+	"strings"
+	"time"
+
+	"github.com/spf13/cobra"
+
+	"gitea.mynodes.xyz/ilker/llmgw/internal/control/adminapi"
+	"gitea.mynodes.xyz/ilker/llmgw/internal/control/catalog"
+	"gitea.mynodes.xyz/ilker/llmgw/internal/control/modelserve"
+	"gitea.mynodes.xyz/ilker/llmgw/internal/guard"
+	"gitea.mynodes.xyz/ilker/llmgw/internal/settings"
+)
+
+func newGuardJudgeCmd(client func() *ctlClient, asJSON *bool) *cobra.Command {
+	c := &cobra.Command{
+		Use:   "judge",
+		Short: "The judge models the policies ask: whether each can answer, and serving one",
+		Long: "A policy with a judge has a model read what patterns cannot. The judge must\n" +
+			"be in the catalog under the name the policy gives, as a judge (role:\n" +
+			"guard), with a replica ready; otherwise what the policy judges is refused\n" +
+			"(fail: closed, guard_unavailable) or served unjudged (fail: open).\n\n" +
+			"A policy may name several, asked in turn (judge.models): a judge on a GPU\n" +
+			"first and one on CPUs after it is the GPU's speed while it is there, and\n" +
+			"the CPUs' when it is not.\n\n" +
+			"Judges llmgw knows where to fetch, which a policy may name and the leader\n" +
+			"then fetches and serves by itself: " + knownJudgeNames() + ".",
+	}
+	c.AddCommand(newGuardJudgeStatusCmd(client, asJSON), newGuardJudgeEnableCmd(client),
+		newGuardJudgeUseCmd(client), newGuardJudgeAskCmd(client, asJSON))
+	return c
+}
+
+func newGuardJudgeStatusCmd(client func() *ctlClient, asJSON *bool) *cobra.Command {
+	return &cobra.Command{
+		Use:   "status",
+		Short: "Each judge the policies ask, and what is wrong with it",
+		Args:  cobra.NoArgs,
+		RunE: func(cmd *cobra.Command, _ []string) error {
+			var js []adminapi.JudgeView
+			if err := client().get(cmd.Context(), "/api/v1/guard/judges", &js); err != nil {
+				return err
+			}
+			if *asJSON {
+				return emitJSON(js)
+			}
+			out := cmd.OutOrStdout()
+			if len(js) == 0 {
+				fmt.Fprintln(out, "no policy asks a judge")
+				return nil
+			}
+			w := tabTo(out)
+			fmt.Fprintln(w, "JUDGE\tPOLICIES\tIN CATALOG\tAS JUDGE\tREADY\tSTATE")
+			for _, j := range js {
+				state := "answering"
+				if j.Problem != "" {
+					state = j.Problem
+				}
+				fmt.Fprintf(w, "%s\t%s\t%s\t%s\t%d/%d\t%s\n", j.Model, strings.Join(j.Policies, ","),
+					yesNo(j.InCatalog), yesNo(j.Judge), j.Ready, j.Wanted, state)
+			}
+			for _, j := range js {
+				if len(j.Then) > 0 {
+					fmt.Fprintf(w, "\n%s gives no verdict: %s is asked instead", j.Model, strings.Join(j.Then, ", else "))
+				}
+			}
+			if err := w.Flush(); err != nil {
+				return err
+			}
+			strays, _ := strayJudges(cmd, client, js)
+			sayStrays(out, strays)
+			return nil
+		},
+	}
+}
+
+// strayJudges are models whose weights are a judge's, under a name no policy
+// asks: a judge the policies moved on from, or one served from the mirror
+// page as a model clients use, as before that page knew better.
+func strayJudges(cmd *cobra.Command, client func() *ctlClient, js []adminapi.JudgeView) ([]adminapi.ModelView, error) {
+	var models []adminapi.ModelView
+	if err := client().get(cmd.Context(), "/api/v1/models", &models); err != nil {
+		return nil, err
+	}
+	asked := map[string]bool{}
+	for _, j := range js {
+		asked[j.Model] = true
+	}
+	var out []adminapi.ModelView
+	for _, m := range models {
+		if _, judge := guard.FormatOfWeights(m.Source); !asked[m.Name] && judge {
+			out = append(out, m)
+		}
+	}
+	return out, nil
+}
+
+// sayStrays says what each stray judge is and how to be rid of it.
+func sayStrays(out io.Writer, strays []adminapi.ModelView) {
+	for _, m := range strays {
+		if m.Role == catalog.RoleGuard {
+			fmt.Fprintf(out, "\n%s is a judge no policy asks, still running: llmgw ctl model remove %s\n", m.Name, m.Name)
+			continue
+		}
+		fmt.Fprintf(out, "\n%s looks like a judge's weights but is served as a model clients use, and no "+
+			"policy asks it: llmgw ctl model remove %s\n", m.Name, m.Name)
+	}
+}
+
+func newGuardJudgeEnableCmd(client func() *ctlClient) *cobra.Command {
+	var (
+		model, repo, file, format, gpuClass string
+		replicas, cores                     int
+		gpu, replace, detach                bool
+	)
+	c := &cobra.Command{
+		Use:   "enable",
+		Short: "Serve a judge: Qwen3Guard-Gen 0.6B on CPUs, or with --gpu the 4B on a GPU",
+		Long: "Fetches a judge's weights into the mirror (once; what is there already, its\n" +
+			"checksums matching, is not fetched again), adds it to the catalog as a judge\n" +
+			"(role: guard) and waits for a replica to answer. Which judge: --model, or\n" +
+			"with --gpu " + defaultGPUJudge + ", or else the first the policies ask that is\n" +
+			"missing, or else " + defaultJudge + ".\n\n" +
+			"A judge on CPUs takes no card. One on a GPU takes one card, as a standard\n" +
+			"model: from nothing more important, so it waits while the cards are busy;\n" +
+			"put a judge on CPUs after it in the policies (llmgw ctl guard judge use) so\n" +
+			"there is always one to ask.\n\n" +
+			"Known judges: " + knownJudgeNames() + ". Another is served with --repo,\n" +
+			"--file and --format. A model of that name already in the catalog as a\n" +
+			"model clients use is refused; --replace removes it first. An offline\n" +
+			"cluster gets its judge from its bundle with server init.",
+		Example: "  llmgw ctl guard judge enable\n" +
+			"  llmgw ctl guard judge enable --gpu\n" +
+			"  llmgw ctl guard judge use " + defaultGPUJudge + " " + defaultJudge,
+		Args: cobra.NoArgs,
+		RunE: func(cmd *cobra.Command, _ []string) error {
+			ctx := cmd.Context()
+			out := cmd.OutOrStdout()
+			var js []adminapi.JudgeView
+			if err := client().get(ctx, "/api/v1/guard/judges", &js); err != nil {
+				return err
+			}
+			name := model
+			if name == "" && gpu {
+				name = defaultGPUJudge
+			}
+			if name == "" {
+				for _, j := range js {
+					if !j.InCatalog || !j.Judge {
+						name = j.Model
+						break
+					}
+				}
+			}
+			if name == "" {
+				name = defaultJudge
+			}
+
+			k, known := knownJudgeNamed(name)
+			switch {
+			case repo != "":
+				f := format
+				if f == "" {
+					f, _ = guard.FormatOfWeights(repo + "/" + file)
+				}
+				if f == "" || !guard.KnownFormat(f) {
+					return fmt.Errorf("say how its answers are read: --format %s", strings.Join(guard.Formats, "|"))
+				}
+				k = knownJudge{Name: name, Repo: repo, File: file, Format: f, GPU: gpu}
+			case !known:
+				return fmt.Errorf("%s is not a judge llmgw knows where to fetch (%s): give --repo, --file and --format", name, knownJudgeNames())
+			}
+			if gpu {
+				k.GPU = true
+			}
+
+			var models []adminapi.ModelView
+			if err := client().get(ctx, "/api/v1/models", &models); err != nil {
+				return err
+			}
+			if i := slices.IndexFunc(models, func(m adminapi.ModelView) bool { return m.Name == name }); i >= 0 {
+				if models[i].Role == catalog.RoleGuard {
+					fmt.Fprintf(out, "%s is in the catalog as a judge already; llmgw ctl guard judge status says how it is\n", name)
+					return nil
+				}
+				if !replace {
+					return fmt.Errorf("%s is in the catalog as a model clients use, not as a judge: "+
+						"--replace removes it and serves it again as a judge", name)
+				}
+				var res map[string]string
+				if err := client().delete(ctx, "/api/v1/models/"+name, &res); err != nil {
+					return fmt.Errorf("removing %s: %w", name, err)
+				}
+				fmt.Fprintf(out, "%s removed\n", name)
+			}
+
+			opts := judgeServeOptions{Replicas: replicas, Cores: cores, GPUClass: gpuClass}
+			if k.GPU && opts.GPUClass == "" {
+				var s adminapi.Summary
+				if err := client().get(ctx, "/api/v1/summary", &s); err != nil {
+					return err
+				}
+				if opts.GPUClass = roomiestGPUClass(s.GPUs); opts.GPUClass == "" {
+					return fmt.Errorf("this cluster has no GPU for %s: serve %s on CPUs instead", name, defaultJudge)
+				}
+			}
+			fmt.Fprintf(out, "serving %s\n", k)
+			body, _ := json.Marshal(k.serveRequest(opts))
+			var job modelserve.Job
+			if err := client().patient(30*time.Second).post(ctx, "/api/v1/models/serve", body, &job); err != nil {
+				return err
+			}
+			if detach {
+				return detached(cmd, job)
+			}
+			if err := followJob(cmd, client, job, "is the judge"); err != nil {
+				return err
+			}
+			if !slices.ContainsFunc(js, func(j adminapi.JudgeView) bool { return j.Model == name }) {
+				fmt.Fprintf(out, "\nno policy asks %s yet: llmgw ctl guard judge use %s", name, name)
+				if k.GPU {
+					fmt.Fprintf(out, " %s", defaultJudge)
+				}
+				fmt.Fprintln(out)
+			}
+			if strays, _ := strayJudges(cmd, client, js); len(strays) > 0 {
+				sayStrays(out, slices.DeleteFunc(strays, func(m adminapi.ModelView) bool { return m.Name == name }))
+			}
+			return nil
+		},
+	}
+	f := c.Flags()
+	f.StringVar(&model, "model", "", "the judge to serve: "+knownJudgeNames()+", or a name for --repo's")
+	f.BoolVar(&gpu, "gpu", false, "serve it on a GPU ("+defaultGPUJudge+" when no --model)")
+	f.StringVar(&gpuClass, "gpu-class", "", "the GPUs it runs on (default: the class with the most free cards)")
+	f.StringVar(&repo, "repo", "", "the registry repository of a judge llmgw does not know")
+	f.StringVar(&file, "file", "", "the GGUF file in --repo")
+	f.StringVar(&format, "format", "", "how --repo's judge answers: "+strings.Join(guard.Formats, ", "))
+	f.IntVar(&replicas, "replicas", 0, "replicas (default: 2 on CPUs, as every judged request waits on one; 1 on a GPU)")
+	f.IntVar(&cores, "cores", 4, "CPU cores a replica on CPUs computes on")
+	f.BoolVar(&replace, "replace", false, "remove a model of that name that is not a judge first")
+	f.BoolVar(&detach, "detach", false, "start it and return; llmgw ctl jobs follows it")
+	return c
+}
+
+// roomiestGPUClass is the class with the most free cards, else the most
+// cards; empty when the cluster has none.
+func roomiestGPUClass(classes []adminapi.ClassUse) string {
+	best := -1
+	for i, c := range classes {
+		if c.Total == 0 {
+			continue
+		}
+		if best < 0 || c.Free() > classes[best].Free() || (c.Free() == classes[best].Free() && c.Total > classes[best].Total) {
+			best = i
+		}
+	}
+	if best < 0 {
+		return ""
+	}
+	return classes[best].Class
+}
+
+func newGuardJudgeUseCmd(client func() *ctlClient) *cobra.Command {
+	var policies []string
+	var note string
+	c := &cobra.Command{
+		Use:   "use JUDGE...",
+		Short: "Set the judges the policies ask, in the order they are asked",
+		Long: "Sets judge.models of every policy that has a judge, or of those given with\n" +
+			"--policy: the first judge that answers gives the verdict, and the next is\n" +
+			"asked when one has no replica serving, fails, or does not answer within\n" +
+			"the policy's latency_budget. A judge on a GPU first and one on CPUs after\n" +
+			"it is the GPU's speed while it is there, and the CPUs' when it is not.\n\n" +
+			"A judge llmgw knows (" + knownJudgeNames() + ")\n" +
+			"that is not in the catalog is fetched and served by the leader; another\n" +
+			"is served first with llmgw ctl guard judge enable --repo. In force on\n" +
+			"every control plane within seconds.",
+		Example: "  llmgw ctl guard judge use " + defaultJudge + "\n" +
+			"  llmgw ctl guard judge use " + defaultGPUJudge + " " + defaultJudge + "\n" +
+			"  llmgw ctl guard judge use " + defaultGPUJudge + " " + defaultJudge + " --policy strict",
+		Args: cobra.MinimumNArgs(1),
+		RunE: func(cmd *cobra.Command, args []string) error {
+			for i, a := range args {
+				if slices.Contains(args[:i], a) {
+					return fmt.Errorf("%s is named twice", a)
+				}
+			}
+			if note == "" {
+				note = "judges: " + strings.Join(args, ", else ")
+			}
+			list := "[" + strings.Join(args, ", ") + "]"
+			return changeSettings(cmd, client, note, func(doc *settings.Doc) error {
+				names := policies
+				if len(names) == 0 {
+					for _, p := range doc.MapKeys([]string{"guard", "policies"}) {
+						if doc.Node([]string{"guard", "policies", p, "judge"}) != nil {
+							names = append(names, p)
+						}
+					}
+					if len(names) == 0 {
+						return errors.New("no policy has a judge: name one with --policy, and it is given one " +
+							"that blocks what the judge finds unsafe in requests (judge.input: block)")
+					}
+				}
+				for _, p := range names {
+					base := []string{"guard", "policies", p}
+					if doc.Node(base) == nil {
+						return fmt.Errorf("there is no policy %s (llmgw ctl guard list)", p)
+					}
+					judge := append(slices.Clone(base), "judge")
+					if doc.Node(judge) == nil {
+						if err := doc.SetYAML(append(slices.Clone(judge), "input"), "block"); err != nil {
+							return err
+						}
+					}
+					doc.Unset(append(slices.Clone(judge), "model"))
+					if err := doc.SetYAML(append(slices.Clone(judge), "models"), list); err != nil {
+						return err
+					}
+					cats := append(slices.Clone(judge), "categories")
+					var have []string
+					if n := doc.Node(cats); n != nil {
+						for _, c := range n.Content {
+							have = append(have, c.Value)
+						}
+					}
+					if next := categoriesFor(args, have); next != nil {
+						if err := doc.SetYAML(cats, "["+strings.Join(next, ", ")+"]"); err != nil {
+							return err
+						}
+						fmt.Fprintf(cmd.OutOrStdout(), "%s: categories [%s] are not what %s name; it counts [%s] instead\n",
+							p, strings.Join(have, ", "), strings.Join(args, " or "), strings.Join(next, ", "))
+					}
+				}
+				return nil
+			})
+		},
+	}
+	c.Flags().StringSliceVar(&policies, "policy", nil, "the policies to change (default: every one with a judge)")
+	c.Flags().StringVar(&note, "note", "", "why, kept with the settings' version")
+	return c
+}
+
+func newGuardJudgeAskCmd(client func() *ctlClient, asJSON *bool) *cobra.Command {
+	var model, answer string
+	c := &cobra.Command{
+		Use:   "ask TEXT",
+		Short: "Have a judge read a text and say what it answered",
+		Long: "For finding out why a request was refused: the judge reads TEXT as a user's\n" +
+			"message (with --answer, ANSWER as the reply to it), past the cache, and its\n" +
+			"answer is shown as it gave it and as the gateway reads it — \"unsafe\" and\n" +
+			"the hazards it named, S6 Specialized Advice and the like. Nothing is kept:\n" +
+			"not in the audit, the usage or the logs.",
+		Example: "  llmgw ctl guard judge ask 'How do I sort a list in Python?'\n" +
+			"  llmgw ctl guard judge ask 'Write a poem' --answer 'Roses are red'",
+		Args: cobra.ExactArgs(1),
+		RunE: func(cmd *cobra.Command, args []string) error {
+			body, _ := json.Marshal(adminapi.JudgeAskRequest{Model: model, Text: args[0], Answer: answer})
+			var v adminapi.JudgeAskView
+			if err := client().patient(45*time.Second).post(cmd.Context(), "/api/v1/guard/judges/ask", body, &v); err != nil {
+				return err
+			}
+			if *asJSON {
+				return emitJSON(v)
+			}
+			out := cmd.OutOrStdout()
+			fmt.Fprintf(out, "%s: %s", v.Model, v.Verdict)
+			if len(v.Categories) > 0 {
+				fmt.Fprintf(out, " (%s)", strings.Join(v.Categories, ", "))
+			}
+			fmt.Fprintf(out, ", %d ms\nit answered: %q\n", v.TookMs, v.Raw)
+			return nil
+		},
+	}
+	c.Flags().StringVar(&model, "model", "", "the judge (default: the one the policies ask)")
+	c.Flags().StringVar(&answer, "answer", "", "an assistant's reply to TEXT, judged as output is")
+	return c
+}

internal/cli/judgekeeper.go

@@ -0,0 +1,116 @@
+package cli
+
+import (
+	"context"
+	"log/slog"
+	"slices"
+	"time"
+
+	"gitea.mynodes.xyz/ilker/llmgw/internal/control/adminapi"
+	"gitea.mynodes.xyz/ilker/llmgw/internal/control/catalog"
+	"gitea.mynodes.xyz/ilker/llmgw/internal/control/modelserve"
+	"gitea.mynodes.xyz/ilker/llmgw/internal/guard"
+)
+
+// judgeKeeper fetches and serves, on the leader, each judge llmgw knows that
+// a policy names and the catalog does not have: naming qwen3guard-4b in a
+// policy is all it takes to have one, and a judge removed by mistake comes
+// back. A judge it does not know is left to llmgw ctl guard judge enable
+// --repo, said in guard judge status.
+type judgeKeeper struct {
+	library func() guard.Config
+	models  func() []catalog.Model
+	serve   interface {
+		Start(modelserve.Request) (modelserve.Job, error)
+		Jobs() []modelserve.Job
+	}
+	leader func() bool
+	gpus   func() []adminapi.ClassUse
+	log    *slog.Logger
+	now    func() time.Time
+
+	// tried is when each judge was last asked for, so one that cannot be
+	// fetched is not asked for again on every look.
+	tried map[string]time.Time
+}
+
+// judgeRetry is how long a judge whose serving failed waits before it is
+// asked for again.
+const judgeRetry = 10 * time.Minute
+
+func (k *judgeKeeper) run(ctx context.Context, every time.Duration) {
+	t := time.NewTicker(every)
+	defer t.Stop()
+	for {
+		k.check()
+		select {
+		case <-ctx.Done():
+			return
+		case <-t.C:
+		}
+	}
+}
+
+func (k *judgeKeeper) check() {
+	if k.leader != nil && !k.leader() {
+		return
+	}
+	lib := k.library()
+	if !lib.On() {
+		return
+	}
+	if k.tried == nil {
+		k.tried = map[string]time.Time{}
+	}
+	now := time.Now
+	if k.now != nil {
+		now = k.now
+	}
+	have := map[string]bool{}
+	for _, m := range k.models() {
+		have[m.Name] = true
+	}
+	busy := map[string]bool{}
+	for _, j := range k.serve.Jobs() {
+		if j.State == modelserve.Running {
+			busy[j.Model] = true
+			busy[j.Request.Name] = true
+		}
+	}
+	var wanted []string
+	for _, p := range lib.Policies {
+		for _, j := range p.Judge.Judges() {
+			if !slices.Contains(wanted, j) {
+				wanted = append(wanted, j)
+			}
+		}
+	}
+	slices.Sort(wanted)
+	for _, name := range wanted {
+		kj, known := knownJudgeNamed(name)
+		if have[name] || !known || busy[name] {
+			continue
+		}
+		if at, ok := k.tried[name]; ok && now().Sub(at) < judgeRetry {
+			continue
+		}
+		k.tried[name] = now()
+		var opts judgeServeOptions
+		if kj.GPU {
+			if opts.GPUClass = roomiestGPUClass(k.gpus()); opts.GPUClass == "" {
+				k.log.Warn("a policy names a judge that runs on a GPU, and this cluster has none",
+					"judge", name, "instead", defaultJudge)
+				continue
+			}
+		}
+		req := kj.serveRequest(opts)
+		req.By = "llmgw (a guard policy names it)"
+		job, err := k.serve.Start(req)
+		if err != nil {
+			k.log.Error("could not start serving a judge a policy names", "judge", name, "err", err)
+			continue
+		}
+		k.log.Info("serving a judge a policy names, which the catalog did not have",
+			"judge", name, "repo", kj.Repo, "file", kj.File, "job", job.ID)
+	}
+}

internal/cli/judgekeeper_test.go

@@ -0,0 +1,126 @@
+package cli
+
+import (
+	"errors"
+	"log/slog"
+	"slices"
+	"testing"
+	"time"
+
+	"gitea.mynodes.xyz/ilker/llmgw/internal/control/adminapi"
+	"gitea.mynodes.xyz/ilker/llmgw/internal/control/catalog"
+	"gitea.mynodes.xyz/ilker/llmgw/internal/control/modelserve"
+	"gitea.mynodes.xyz/ilker/llmgw/internal/guard"
+)
+
+type fakeServe struct {
+	started []modelserve.Request
+	running []modelserve.Job
+	err     error
+}
+
+func (f *fakeServe) Start(r modelserve.Request) (modelserve.Job, error) {
+	if f.err != nil {
+		return modelserve.Job{}, f.err
+	}
+	f.started = append(f.started, r)
+	return modelserve.Job{ID: "j", Model: r.Name, Request: r, State: modelserve.Running}, nil
+}
+func (f *fakeServe) Jobs() []modelserve.Job { return f.running }
+
+func TestAJudgeAPolicyNamesIsFetchedWhenTheCatalogLacksIt(t *testing.T) {
+	lib := guard.Config{Policies: map[string]guard.Policy{
+		"baseline": {Judge: &guard.JudgePolicy{Models: []string{"qwen3guard-4b", "qwen3guard-0.6b"}, Input: "block"}},
+		"custom":   {Judge: &guard.JudgePolicy{Model: "our-own-judge", Input: "block"}},
+	}}
+	inCatalog := []catalog.Model{{Name: "qwen3guard-0.6b"}}
+	now := time.Date(2026, 10, 8, 12, 0, 0, 0, time.UTC)
+	fs := &fakeServe{}
+	leader := true
+	k := &judgeKeeper{
+		library: func() guard.Config { return lib },
+		models:  func() []catalog.Model { return inCatalog },
+		serve:   fs,
+		leader:  func() bool { return leader },
+		gpus:    func() []adminapi.ClassUse { return []adminapi.ClassUse{{Class: "6000-24g", Total: 1, Used: 1}} },
+		log:     slog.New(slog.DiscardHandler),
+		now:     func() time.Time { return now },
+	}
+
+	k.check()
+	// The GPU judge alone: the CPU one is there, and one llmgw does not know
+	// is not guessed at.
+	if len(fs.started) != 1 || fs.started[0].Name != "qwen3guard-4b" || fs.started[0].GPUClass != "6000-24g" ||
+		fs.started[0].Class != catalog.ClassStandard || fs.started[0].GuardFormat != "qwen3guard" || fs.started[0].CPU {
+		t.Fatalf("started %+v", fs.started)
+	}
+
+	// Being fetched: not asked for again.
+	fs.running = []modelserve.Job{{Model: "qwen3guard-4b", State: modelserve.Running}}
+	k.check()
+	if len(fs.started) != 1 {
+		t.Errorf("asked for again while it is being fetched: %d", len(fs.started))
+	}
+	// Failed: asked for again only after a while.
+	fs.running = nil
+	now = now.Add(time.Minute)
+	k.check()
+	if len(fs.started) != 1 {
+		t.Error("asked for again at once after it failed")
+	}
+	now = now.Add(judgeRetry)
+	k.check()
+	if len(fs.started) != 2 {
+		t.Error("never asked for again")
+	}
+
+	// Only the leader fetches, and nothing while the guardrails are off.
+	leader = false
+	now = now.Add(judgeRetry)
+	k.check()
+	leader = true
+	off := false
+	lib.Enabled = &off
+	k.check()
+	if len(fs.started) != 2 {
+		t.Errorf("a standby or the guardrails off fetched: %d", len(fs.started))
+	}
+
+	// A CPU judge on a cluster without GPUs is served on CPUs, two of it.
+	lib = guard.Config{Policies: map[string]guard.Policy{"b": {Judge: &guard.JudgePolicy{Model: "qwen3guard-0.6b", Input: "block"}}}}
+	inCatalog = nil
+	k.gpus = func() []adminapi.ClassUse { return nil }
+	k.check()
+	if last := fs.started[len(fs.started)-1]; last.Name != "qwen3guard-0.6b" || !last.CPU || last.Replicas != 2 {
+		t.Errorf("CPU judge: %+v", last)
+	}
+	fs.err = errors.New("the mirror is down")
+	lib = guard.Config{Policies: map[string]guard.Policy{"b": {Judge: &guard.JudgePolicy{Model: "qwen3guard-8b", Input: "block"}}}}
+	k.check() // no GPU, so not started, and no panic on the error path either
+}
+
+// A policy moved from Llama Guard's codes to Qwen3Guard would count nothing:
+// its categories become the judge's own.
+func TestCategoriesFollowTheJudges(t *testing.T) {
+	lg := []string{"S1", "S2", "S10"}
+	if got := categoriesFor([]string{"qwen3guard-0.6b"}, lg); !slices.Equal(got, baselineCategories["qwen3guard"]) {
+		t.Errorf("Llama Guard's codes for Qwen3Guard: %v", got)
+	}
+	if got := categoriesFor([]string{"qwen3guard-4b", "qwen3guard-0.6b"}, []string{"Violent", "Jailbreak"}); got != nil {
+		t.Errorf("its own names were replaced: %v", got)
+	}
+	if got := categoriesFor([]string{"llama-guard-3-1b"}, lg); got != nil {
+		t.Errorf("Llama Guard's own codes were replaced: %v", got)
+	}
+	// Both kinds asked in turn: each must have something to count.
+	got := categoriesFor([]string{"qwen3guard-4b", "llama-guard-3-1b"}, []string{"Violent"})
+	if !slices.Contains(got, "Violent") || !slices.Contains(got, "S1") {
+		t.Errorf("mixed: %v", got)
+	}
+	if got := categoriesFor([]string{"our-own-judge"}, lg); got != nil {
+		t.Errorf("a judge llmgw does not know: %v", got)
+	}
+	if got := categoriesFor([]string{"qwen3guard-0.6b"}, nil); got != nil {
+		t.Errorf("no categories (every one counts) became %v", got)
+	}
+}

internal/cli/judges_known.go

@@ -0,0 +1,158 @@
+package cli
+
+import (
+	"fmt"
+	"slices"
+	"strings"
+
+	"gitea.mynodes.xyz/ilker/llmgw/internal/control/catalog"
+	"gitea.mynodes.xyz/ilker/llmgw/internal/control/modelserve"
+	"gitea.mynodes.xyz/ilker/llmgw/internal/guard"
+)
+
+// knownJudge is a judge model llmgw knows where to fetch and how to serve:
+// what a policy may name without anyone saying where its weights are.
+type knownJudge struct {
+	Name, Repo, File, Format string
+	// GPU is a judge too slow on CPUs to wait for: served on one card.
+	GPU   bool
+	About string
+}
+
+// knownJudges, the default first. Measured on the reference cluster against
+// ordinary coding requests ("continue", "kill the process on port 8080") and
+// harmful ones: Llama Guard 3 1B scored both alike, about one in two
+// ordinary ones refused; Qwen3Guard-Gen 0.6B kept them apart, 0.15 against
+// 1.00 at the worst, at about two seconds a check on two CPU threads.
+var knownJudges = []knownJudge{
+	{Name: "qwen3guard-0.6b", Repo: "mradermacher/Qwen3Guard-Gen-0.6B-GGUF", File: "Qwen3Guard-Gen-0.6B.Q8_0.gguf",
+		Format: "qwen3guard", About: "Qwen3Guard-Gen 0.6B on CPUs, 0.8 GB: the default"},
+	{Name: "qwen3guard-4b", Repo: "mradermacher/Qwen3Guard-Gen-4B-GGUF", File: "Qwen3Guard-Gen-4B.Q4_K_M.gguf",
+		Format: "qwen3guard", GPU: true, About: "Qwen3Guard-Gen 4B on a GPU, 2.5 GB"},
+	{Name: "qwen3guard-8b", Repo: "mradermacher/Qwen3Guard-Gen-8B-GGUF", File: "Qwen3Guard-Gen-8B.Q4_K_M.gguf",
+		Format: "qwen3guard", GPU: true, About: "Qwen3Guard-Gen 8B on a GPU, 4.7 GB"},
+	{Name: "llama-guard-3-1b", Repo: "QuantFactory/Llama-Guard-3-1B-GGUF", File: "Llama-Guard-3-1B.Q4_K_M.gguf",
+		Format: "llamaguard3", About: "Llama Guard 3 1B on CPUs, 0.9 GB: refuses ordinary coding requests"},
+	{Name: "llama-guard-3-8b", Repo: "QuantFactory/Llama-Guard-3-8B-GGUF", File: "Llama-Guard-3-8B.Q4_K_M.gguf",
+		Format: "llamaguard3", GPU: true, About: "Llama Guard 3 8B on a GPU, 4.6 GB"},
+}
+
+const (
+	defaultJudge    = "qwen3guard-0.6b"
+	defaultGPUJudge = "qwen3guard-4b"
+)
+
+// baselineCategories are what the baseline policy counts of each format's
+// verdicts: not PII, which is the patterns', nor what refuses ordinary work
+// — copyright, unethical acts, politically sensitive topics; Llama Guard's
+// defamation, specialized advice, intellectual property and elections.
+var baselineCategories = map[string][]string{
+	"qwen3guard":  {"Violent", "Non-violent Illegal Acts", "Sexual Content or Sexual Acts", "Suicide & Self-Harm", "Jailbreak"},
+	"llamaguard3": {"S1", "S2", "S3", "S4", "S7", "S9", "S10", "S11", "S12"},
+}
+
+// reported is whether a judge of that format ever names category c: its own
+// names, or for Llama Guard its codes.
+func reported(format, c string) bool {
+	if slices.ContainsFunc(guard.DefaultCategories(format), func(n string) bool { return strings.EqualFold(n, c) }) {
+		return true
+	}
+	var n int
+	_, err := fmt.Sscanf(c, "S%d", &n)
+	return format == "llamaguard3" && err == nil && n >= 1 && n <= 14
+}
+
+// categoriesFor is what a policy asking these judges should count, when the
+// categories it has are ones none of them names: a policy moved from Llama
+// Guard's S1 to Qwen3Guard, which says Violent, would otherwise count
+// nothing. Nil keeps what it has; so does a judge llmgw does not know.
+func categoriesFor(judges, have []string) []string {
+	var formats []string
+	for _, j := range judges {
+		k, ok := knownJudgeNamed(j)
+		if !ok {
+			return nil
+		}
+		if !slices.Contains(formats, k.Format) {
+			formats = append(formats, k.Format)
+		}
+	}
+	if len(have) == 0 {
+		return nil
+	}
+	stray := false
+	for _, f := range formats {
+		for _, c := range have {
+			if !reported(f, c) && !slices.ContainsFunc(formats, func(o string) bool { return reported(o, c) }) {
+				stray = true
+			}
+		}
+		// Every judge must have something it counts.
+		if !slices.ContainsFunc(have, func(c string) bool { return reported(f, c) }) {
+			stray = true
+		}
+	}
+	if !stray {
+		return nil
+	}
+	var out []string
+	for _, f := range formats {
+		out = append(out, baselineCategories[f]...)
+	}
+	return out
+}
+
+func knownJudgeNamed(name string) (knownJudge, bool) {
+	i := slices.IndexFunc(knownJudges, func(k knownJudge) bool { return k.Name == name })
+	if i < 0 {
+		return knownJudge{}, false
+	}
+	return knownJudges[i], true
+}
+
+func knownJudgeNames() string {
+	var n []string
+	for _, k := range knownJudges {
+		n = append(n, k.Name)
+	}
+	return strings.Join(n, ", ")
+}
+
+// judgeServeOptions are how a judge is served beyond which weights.
+type judgeServeOptions struct {
+	Replicas, Cores int
+	// GPUClass is the cards a GPU judge runs on; empty lets the plan choose.
+	GPUClass string
+}
+
+// serveRequest is the serve job for a judge. A judge on CPUs is critical,
+// as every judged request waits on it; one on a GPU is standard, so it takes
+// a card only from what may give one up, and the judge on CPUs after it in
+// the policies stands in while it has none.
+func (k knownJudge) serveRequest(o judgeServeOptions) modelserve.Request {
+	req := modelserve.Request{
+		Repo: k.Repo, File: k.File, Name: k.Name, GuardFormat: k.Format,
+		Description: "The guardrails' judge: reads prompts and answers for unsafe content. Not offered to clients.",
+		Replicas:    o.Replicas, CtxPool: 8192, ContextLength: 4096,
+	}
+	if k.GPU {
+		req.GPUClass, req.GPUsPerNode, req.Class = o.GPUClass, 1, catalog.ClassStandard
+		if req.Replicas == 0 {
+			req.Replicas = 1
+		}
+		return req
+	}
+	req.CPU, req.CPUCores = true, o.Cores
+	if req.Replicas == 0 {
+		req.Replicas = 2
+	}
+	return req
+}
+
+func (k knownJudge) String() string {
+	where := "CPUs"
+	if k.GPU {
+		where = "a GPU"
+	}
+	return fmt.Sprintf("%s (%s, %s/%s, on %s)", k.Name, k.Format, k.Repo, k.File, where)
+}

internal/cli/kb.go

@@ -0,0 +1,332 @@
+package cli
+
+import (
+	"bytes"
+	"encoding/json"
+	"fmt"
+	"io"
+	"net/url"
+	"os"
+	"strings"
+	"time"
+
+	"github.com/spf13/cobra"
+
+	"gitea.mynodes.xyz/ilker/llmgw/internal/control/adminapi"
+)
+
+func newKBCmd(client func() *ctlClient, asJSON *bool) *cobra.Command {
+	c := &cobra.Command{
+		Use:   "kb",
+		Short: "Knowledge bases: collections of documents requests are answered from",
+		Long: "A collection belongs to an organization, team, project or person, and is\n" +
+			"embedded by one model of the catalog. People put documents into it from\n" +
+			"their own machines with llmgw-kb (its own download, for Windows, macOS and\n" +
+			"Linux), which sends text and vectors, never the files.\n\n" +
+			"Who may read or add to a collection: its owner and everything below it,\n" +
+			"and whoever it is granted to (grant), never outside its organization.\n" +
+			"Targets are written org:Acme, team:Acme/hr, project:Acme/hr/docs (or with\n" +
+			"IDs), key:KEY-ID or user:USER-ID. A collection is named by its ID, its name\n" +
+			"when that is unique, or ORG/NAME.\n\n" +
+			"Needs the PostgreSQL store with pgvector; `llmgw ctl kb list` says why not.",
+	}
+	c.AddCommand(&cobra.Command{
+		Use:   "list",
+		Short: "Every collection: owner, model, documents and grants",
+		Args:  cobra.NoArgs,
+		RunE: func(cmd *cobra.Command, _ []string) error {
+			var v adminapi.KBView
+			if err := client().get(cmd.Context(), "/api/v1/kb", &v); err != nil {
+				return err
+			}
+			if *asJSON {
+				return json.NewEncoder(cmd.OutOrStdout()).Encode(v)
+			}
+			out := cmd.OutOrStdout()
+			if !v.Available {
+				fmt.Fprintln(out, "Knowledge bases are off: "+v.Reason)
+				return nil
+			}
+			if len(v.Collections) == 0 {
+				fmt.Fprintln(out, "No collections yet (llmgw ctl kb create).")
+				return nil
+			}
+			w := tab()
+			fmt.Fprintln(w, "COLLECTION\tOWNER\tMODEL\tDOCUMENTS\tCHUNKS\tGRANTED TO")
+			for _, c := range v.Collections {
+				var grants []string
+				for _, g := range c.Grants {
+					grants = append(grants, or(g.Name, g.Target)+" ("+g.Perm+")")
+				}
+				fmt.Fprintf(w, "%s/%s\t%s\t%s (%d)\t%d\t%d\t%s\n", or(c.OrgName, c.Org), c.Name, or(c.OwnerName, c.Owner),
+					c.EmbedModel, c.Dims, c.Documents, c.Chunks, strings.Join(grants, ", "))
+			}
+			if err := w.Flush(); err != nil {
+				return err
+			}
+			if len(v.Defaults) > 0 {
+				fmt.Fprintln(out, "\nAnswered from by default, when a request names no collection:")
+				w = tab()
+				for _, d := range v.Defaults {
+					fmt.Fprintf(w, "  %s\t%s\n", or(d.Name, d.Target), strings.Join(d.Collections, ", "))
+				}
+				return w.Flush()
+			}
+			return nil
+		},
+	})
+
+	var org string
+	var dims int
+	create := &cobra.Command{
+		Use:   "create OWNER NAME MODEL",
+		Short: "Make a collection owned by an organization, team, project or person",
+		Long: "MODEL is a catalog model with capabilities.embeddings. Its dimension is asked\n" +
+			"of it, so it has to be able to serve now; --dims gives it instead.",
+		Example: "  llmgw ctl kb create team:Acme/hr policies bge-m3\n" +
+			"  llmgw ctl kb create user:USER-ID notes bge-m3 --org Acme",
+		Args: cobra.ExactArgs(3),
+		RunE: func(cmd *cobra.Command, args []string) error {
+			body, _ := json.Marshal(adminapi.KBCreateRequest{Owner: args[0], Name: args[1], Model: args[2], Org: org, Dims: dims})
+			var v adminapi.KBCollectionView
+			if err := client().do(cmd.Context(), "POST", "/api/v1/kb/collections", json.RawMessage(body), &v); err != nil {
+				return err
+			}
+			fmt.Fprintf(cmd.OutOrStdout(), "collection %s (%s), embedded by %s in %d dimensions\n", v.Name, v.ID, v.EmbedModel, v.Dims)
+			fmt.Fprintf(cmd.OutOrStdout(), "people with access fill it with: llmgw-kb sync %s FOLDER\n", v.Name)
+			return nil
+		},
+	}
+	create.Flags().StringVar(&org, "org", "", "the organization of a person's collection")
+	create.Flags().IntVar(&dims, "dims", 0, "the model's embedding dimension, instead of asking it")
+	c.AddCommand(create)
+
+	c.AddCommand(&cobra.Command{
+		Use:   "delete COLLECTION",
+		Short: "Remove a collection, its documents and its grants",
+		Args:  cobra.ExactArgs(1),
+		RunE: func(cmd *cobra.Command, args []string) error {
+			if err := client().do(cmd.Context(), "DELETE", "/api/v1/kb/collections/"+url.PathEscape(args[0]), nil, nil); err != nil {
+				return err
+			}
+			fmt.Fprintf(cmd.OutOrStdout(), "collection %s removed\n", args[0])
+			return nil
+		},
+	})
+
+	for _, verb := range []struct {
+		use, short, method string
+		perm               bool
+	}{
+		{"grant", "Let a scope and everything below it read a collection, or with --write add to it", "POST", true},
+		{"revoke", "Take a grant away", "DELETE", false},
+	} {
+		var write bool
+		cmd := &cobra.Command{
+			Use:     verb.use + " COLLECTION TARGET",
+			Short:   verb.short,
+			Example: "  llmgw ctl kb " + verb.use + " Acme/policies project:Acme/sales/crm",
+			Args:    cobra.ExactArgs(2),
+			RunE: func(cmd *cobra.Command, args []string) error {
+				req := adminapi.KBGrantRequest{Collection: args[0], Target: args[1], Perm: "read"}
+				if write {
+					req.Perm = "write"
+				}
+				body, _ := json.Marshal(req)
+				if err := client().do(cmd.Context(), verb.method, "/api/v1/kb/grants", json.RawMessage(body), nil); err != nil {
+					return err
+				}
+				if verb.perm {
+					fmt.Fprintf(cmd.OutOrStdout(), "%s may %s %s\n", args[1], req.Perm, args[0])
+				} else {
+					fmt.Fprintf(cmd.OutOrStdout(), "%s no longer granted %s\n", args[1], args[0])
+				}
+				return nil
+			},
+		}
+		if verb.perm {
+			cmd.Flags().BoolVar(&write, "write", false, "add and remove documents too")
+		}
+		c.AddCommand(cmd)
+	}
+
+	c.AddCommand(&cobra.Command{
+		Use:   "docs COLLECTION",
+		Short: "A collection's documents",
+		Args:  cobra.ExactArgs(1),
+		RunE: func(cmd *cobra.Command, args []string) error {
+			var docs []adminapi.KBDocumentView
+			if err := client().get(cmd.Context(), "/api/v1/kb/collections/"+url.PathEscape(args[0])+"/documents", &docs); err != nil {
+				return err
+			}
+			if *asJSON {
+				return json.NewEncoder(cmd.OutOrStdout()).Encode(docs)
+			}
+			w := tab()
+			fmt.Fprintln(w, "DOCUMENT\tCHUNKS\tSIZE\tADDED\tBY")
+			for _, d := range docs {
+				fmt.Fprintf(w, "%s\t%d\t%s\t%s\t%s\n", d.Name, d.Chunks, humanBytes(d.Bytes), d.AddedAt.Local().Format("2006-01-02 15:04"), d.AddedBy)
+			}
+			return w.Flush()
+		},
+	})
+	c.AddCommand(&cobra.Command{
+		Use:   "rm COLLECTION DOCUMENT",
+		Short: "Remove one document from a collection",
+		Args:  cobra.ExactArgs(2),
+		RunE: func(cmd *cobra.Command, args []string) error {
+			path := "/api/v1/kb/collections/" + url.PathEscape(args[0]) + "/documents?name=" + url.QueryEscape(args[1])
+			if err := client().do(cmd.Context(), "DELETE", path, nil, nil); err != nil {
+				return err
+			}
+			fmt.Fprintf(cmd.OutOrStdout(), "%s removed from %s\n", args[1], args[0])
+			return nil
+		},
+	})
+
+	var clear bool
+	def := &cobra.Command{
+		Use:   "default TARGET [COLLECTION...]",
+		Short: "Answer a part of the tree's chat requests from collections, when they name none",
+		Long: "Requests from TARGET and everything below it are then answered from these\n" +
+			"collections without naming them (X-Llmgw-Knowledge). The narrowest default\n" +
+			"wins: a key's over its project's, a project's over its team's. A request's own\n" +
+			"header overrides it, and X-Llmgw-Knowledge: none turns it off. A collection a\n" +
+			"caller may not read is left out for them, and a search that fails answers\n" +
+			"without passages rather than refusing. In force within 30 seconds.",
+		Example: "  llmgw ctl kb default team:Acme/hr Acme/policies Acme/forms\n" +
+			"  llmgw ctl kb default team:Acme/hr --clear",
+		Args: cobra.MinimumNArgs(1),
+		RunE: func(cmd *cobra.Command, args []string) error {
+			if clear != (len(args) == 1) {
+				return fmt.Errorf("name collections, or --clear")
+			}
+			body, _ := json.Marshal(adminapi.KBDefaultRequest{Target: args[0], Collections: args[1:]})
+			if err := client().do(cmd.Context(), "POST", "/api/v1/kb/defaults", json.RawMessage(body), nil); err != nil {
+				return err
+			}
+			if clear {
+				fmt.Fprintf(cmd.OutOrStdout(), "%s answers from no collection by default\n", args[0])
+			} else {
+				fmt.Fprintf(cmd.OutOrStdout(), "%s answers from %s by default\n", args[0], strings.Join(args[1:], ", "))
+			}
+			return nil
+		},
+	}
+	def.Flags().BoolVar(&clear, "clear", false, "no default any more")
+	c.AddCommand(def)
+
+	var dsModel, dsOut string
+	var dsPer, dsMax int
+	dataset := &cobra.Command{
+		Use:   "dataset COLLECTION --model MODEL",
+		Short: "Write training data from a collection: questions and answers, for an adapter",
+		Long: "A chat model of the cluster reads each chunk of the collection and writes\n" +
+			"questions a person might ask of it, with the answers it holds, one JSON object\n" +
+			"a line ({\"q\", \"a\", \"doc\", \"ord\"}). Read it through and correct it before\n" +
+			"training on it (llmgw ctl train): what the model got wrong, an adapter learns.\n\n" +
+			"An adapter learns how answers sound — the organisation's words, format and\n" +
+			"tone — far better than what they say; the facts stay the knowledge base's.",
+		Example: "  llmgw ctl kb dataset Acme/policies --model qwen2.5-coder-14b-instruct --out policies.jsonl",
+		Args:    cobra.ExactArgs(1),
+		RunE: func(cmd *cobra.Command, args []string) error {
+			w := cmd.OutOrStdout()
+			if dsOut != "" {
+				f, err := os.Create(dsOut)
+				if err != nil {
+					return err
+				}
+				defer f.Close()
+				w = f
+			}
+			lw := &datasetWriter{pairs: w, problems: cmd.ErrOrStderr()}
+			body, _ := json.Marshal(adminapi.KBDatasetRequest{Model: dsModel, PerChunk: dsPer, MaxChunks: dsMax})
+			err := client().patient(6*time.Hour).do(cmd.Context(), "POST",
+				"/api/v1/kb/collections/"+url.PathEscape(args[0])+"/dataset", json.RawMessage(body), lw)
+			lw.flush()
+			fmt.Fprintf(cmd.ErrOrStderr(), "%d pairs from %d chunks; %d chunks gave none\n", lw.n, len(lw.chunks), lw.failed)
+			return err
+		},
+	}
+	dataset.Flags().StringVar(&dsModel, "model", "", "the chat model that writes the questions and answers")
+	dataset.Flags().StringVar(&dsOut, "out", "", "the file to write (default: standard output)")
+	dataset.Flags().IntVar(&dsPer, "per-chunk", 8, "pairs from each chunk")
+	dataset.Flags().IntVar(&dsMax, "max-chunks", 500, "the most chunks read")
+	_ = dataset.MarkFlagRequired("model")
+	c.AddCommand(dataset)
+
+	var k int
+	search := &cobra.Command{
+		Use:   "search COLLECTION QUERY...",
+		Short: "Try a search of a collection, as an administrator (no tenancy applies)",
+		Args:  cobra.MinimumNArgs(2),
+		RunE: func(cmd *cobra.Command, args []string) error {
+			body, _ := json.Marshal(adminapi.KBSearchRequest{Collection: args[0], Query: strings.Join(args[1:], " "), K: k})
+			var hits []adminapi.KBHitView
+			if err := client().do(cmd.Context(), "POST", "/api/v1/kb/search", json.RawMessage(body), &hits); err != nil {
+				return err
+			}
+			if *asJSON {
+				return json.NewEncoder(cmd.OutOrStdout()).Encode(hits)
+			}
+			out := cmd.OutOrStdout()
+			if len(hits) == 0 {
+				fmt.Fprintln(out, "Nothing found.")
+			}
+			for i, h := range hits {
+				fmt.Fprintf(out, "%d. %s #%d\n%s\n\n", i+1, h.Document, h.Ord, indent(strings.TrimSpace(h.Text), "   "))
+			}
+			return nil
+		},
+	}
+	search.Flags().IntVarP(&k, "k", "k", 5, "how many chunks")
+	c.AddCommand(search)
+	return c
+}
+
+// datasetWriter splits the dataset stream: pairs to the file, problems to
+// the terminal, counted as they pass.
+type datasetWriter struct {
+	pairs, problems io.Writer
+	buf             []byte
+	n, failed       int
+	chunks          map[string]bool
+}
+
+func (d *datasetWriter) Write(p []byte) (int, error) {
+	d.buf = append(d.buf, p...)
+	for {
+		i := bytes.IndexByte(d.buf, '\n')
+		if i < 0 {
+			return len(p), nil
+		}
+		d.line(d.buf[:i])
+		d.buf = d.buf[i+1:]
+	}
+}
+
+func (d *datasetWriter) flush() {
+	if len(bytes.TrimSpace(d.buf)) > 0 {
+		d.line(d.buf)
+	}
+	d.buf = nil
+}
+
+func (d *datasetWriter) line(b []byte) {
+	var l adminapi.KBDatasetLine
+	if json.Unmarshal(b, &l) != nil {
+		return
+	}
+	if d.chunks == nil {
+		d.chunks = map[string]bool{}
+	}
+	d.chunks[fmt.Sprintf("%s#%d", l.Doc, l.Ord)] = true
+	if l.Error != "" {
+		d.failed++
+		fmt.Fprintf(d.problems, "  %s #%d: %s\n", l.Doc, l.Ord, l.Error)
+		return
+	}
+	d.n++
+	out, _ := json.Marshal(l)
+	fmt.Fprintln(d.pairs, string(out))
+}

internal/cli/mirrorinit.go

@@ -8,6 +8,7 @@ import (
 	"net"
 	"net/netip"
 	"os"
+	"path/filepath"
 	"runtime"
 	"strings"
 	"time"
@@ -65,6 +66,14 @@ Safe to run again: it changes only what differs.`,
 			}
 			if !noFirewall {
 				mirrorFirewall(cmd.Context(), cmd.OutOrStdout(), controlPlanes)
+				// And after, as the cluster's settings say: the control plane
+				// tells the mirror, which writes it beside its models.
+				state := filepath.Join(*dir, mirror.FirewallFile)
+				if err := installFirewallSync(cmd.Context(), firewall.Local{}, state); err != nil {
+					fmt.Fprintf(cmd.OutOrStdout(), "  !!    firewall: %v; the cluster's settings will not change it here\n", err)
+				} else {
+					fmt.Fprintln(cmd.OutOrStdout(), "  ok    firewall: follows the cluster's settings ("+firewallSyncPathUnit+")")
+				}
 			}
 			return nil
 		},

internal/cli/root.go

@@ -63,6 +63,10 @@ func Root() *cobra.Command {
 		newWipeCmd(),
 		newDocsCmd(),
 	)
+	// status, start, stop and restart: the whole system; service: this
+	// machine's services.
+	root.AddCommand(newSystemCmds()...)
+	root.AddCommand(newServiceCmd())
 	return root
 }
 

internal/cli/server.go

@@ -5,6 +5,7 @@ import (
 	"crypto/rand"
 	"crypto/tls"
 	"encoding/hex"
+	"encoding/json"
 	"errors"
 	"fmt"
 	"log/slog"
@@ -48,6 +49,7 @@ import (
 	"gitea.mynodes.xyz/ilker/llmgw/internal/control/fleetmetrics"
 	"gitea.mynodes.xyz/ilker/llmgw/internal/control/modelsvc"
 	"gitea.mynodes.xyz/ilker/llmgw/internal/control/reconciler"
+	"gitea.mynodes.xyz/ilker/llmgw/internal/control/train"
 	"gitea.mynodes.xyz/ilker/llmgw/internal/guard"
 	"gitea.mynodes.xyz/ilker/llmgw/internal/hf"
 	"gitea.mynodes.xyz/ilker/llmgw/internal/kvevents"
@@ -186,6 +188,7 @@ func newServerCmd() *cobra.Command {
 
 			rt.Start(ctx)
 			defer rt.Stop()
+			rt.SetStopped(cfg.Serving.Stopped)
 
 			// Which cluster this is, to federated peers: the id in the hop
 			// header that stops a request going round two clusters that
@@ -230,6 +233,9 @@ func newServerCmd() *cobra.Command {
 			var agentHandler http.Handler
 			var agentPath string
 			var admin *control.Controller
+			// What the settings say of the firewall, for the mirror and the
+			// store's members; nil without a control plane.
+			var fwPub *firewallPublisher
 			updateMetrics := &modelserve.UpdateCollector{}
 			// The store's cluster, when it is one; read by the metrics and
 			// by `ctl status`.
@@ -252,6 +258,8 @@ func newServerCmd() *cobra.Command {
 				}
 				defer ctl.Close()
 				agentSvc = svc
+				// Stopped by the operators (llmgw stop): nothing placed.
+				ctl.Reconciler().SetStopped(cfg.Serving.Stopped)
 				// Whether nodes keep llmgw's firewall, as the settings say now.
 				svc.SetFirewall(func() *llmgwv1.Firewall {
 					c := current.Load()
@@ -259,9 +267,31 @@ func newServerCmd() *cobra.Command {
 					return &llmgwv1.Firewall{Enabled: c.Firewall.Enabled, Networks: c.FirewallNetworks,
 						EnginePortsFrom: int32(from), EnginePortsTo: int32(to)} //nolint:gosec // ports
 				})
+				// What each node's agent runs on, as the settings say now.
+				svc.SetAgentSettings(func(node string) string {
+					b, err := config.Marshal(current.Load().Agents.For(node))
+					if err != nil {
+						log.Warn("could not write a node's agent settings", "node", node, "err", err)
+						return ""
+					}
+					return string(b)
+				})
 				ctl.Reconciler().SetNotifier(notifier)
 				admin = ctl
 				agentPath, agentHandler = svc.Handler()
+				fwPub = newFirewallPublisher(func() config.Server { return *current.Load() }, rt.Store(),
+					func() []string {
+						var out []string
+						for _, n := range ctl.Inventory().Nodes() {
+							out = append(out, n.MgmtAddr)
+						}
+						return out
+					},
+					func() bool { return ctl.Leadership() == nil || ctl.Leadership().IsLeader() }, log)
+				go fwPub.run(ctx, 5*time.Minute)
+				if l := ctl.Leadership(); l != nil {
+					l.OnLead(fwPub.Kick)
+				}
 
 				// Before AdminHandler, which is where the registry is built —
 				// registering after it silently publishes nothing, which is
@@ -361,16 +391,32 @@ func newServerCmd() *cobra.Command {
 				evals = newEvalService(cfg, rt, admin, log)
 				api.SetEval(evals)
 				if st := rt.Store(); st != nil && gate != nil {
-					api.SetGuard(&adminapi.GuardAdmin{Library: gate.Library, Store: st, Invalidate: gate.Invalidate})
+					api.SetGuard(&adminapi.GuardAdmin{Library: gate.Library, Store: st, Invalidate: gate.Invalidate,
+						JudgeErrors: gate.JudgeErrors, Ask: gate.AskJudge})
 				}
 				if st := rt.Store(); st != nil {
 					// Removing a node revokes its certificate.
 					api.SetRevoker(st.RevokeNode)
+					if agentSvc != nil {
+						api.SetAgentRestarter(agentSvc.RestartAgents)
+					}
 					api.SetAudit(st)
+					// Knowledge bases, with the cluster's own models to embed a
+					// probe (a collection's dimension) and an operator's search.
+					api.SetKB(&adminapi.KBAdmin{Store: st, Embed: embedder(rt.Replay)})
+					// LoRA adapters trained on the nodes.
+					if agentSvc != nil {
+						trainer := train.New(st, agentSvc, log)
+						agentSvc.SetTrain(trainer)
+						go trainer.Run(ctx)
+						api.SetTrain(&adminapi.TrainAdmin{Store: st, Service: trainer})
+					}
 					// The cluster's settings, each change checked against this
 					// machine's own keys before it is kept.
 					api.SetSettings(&adminapi.Settings{Store: st, LocalKeys: config.LocalKeys(),
 						Check: func(body []byte) error { return checkClusterSettings(configPath, body) }})
+					// And the same document on the portal, for administrators.
+					rt.SetSettingsCheck(func(body []byte) error { return checkClusterSettings(configPath, body) }, config.LocalKeys())
 				}
 				// Catalog changes need somewhere to persist, so this is wired
 				// only when there is a store. Without one, the endpoints say so
@@ -427,6 +473,7 @@ func newServerCmd() *cobra.Command {
 					api.SetUsage(st)
 					// Capacity quotas, set and read over this API.
 					api.SetQuotas(st)
+					api.SetTenancy(st)
 				}
 				api.Register(adminMux)
 
@@ -488,6 +535,16 @@ func newServerCmd() *cobra.Command {
 					return fmt.Errorf("the model mirror: %w", err)
 				}
 				rt.SetMirror(mc)
+				if fwPub != nil {
+					fwPub.setMirror(mc)
+				}
+				// Adapters trained on the nodes are kept there too; sending one
+				// takes longer than the mirror's usual answers.
+				if adminAPI != nil {
+					if up, err := mirror.NewClient(cfg.Mirror.API, cfg.Mirror.TLS, 10*time.Minute); err == nil {
+						adminAPI.SetTrainMirror(up)
+					}
+				}
 				// Browsing the registry is what turns "paste a resolve URL"
 				// into "find the model". It needs a way out to the internet
 				// from this machine, which the mirror's own fetching does not.
@@ -511,6 +568,21 @@ func newServerCmd() *cobra.Command {
 					})
 					go serve.WatchCanaries(ctx, 30*time.Second)
 					adminAPI.SetServe(serve)
+					if gate != nil {
+						// A judge a policy names and the catalog lacks is
+						// fetched and served, on the leader.
+						keeper := &judgeKeeper{
+							library: gate.Library,
+							models:  func() []catalog.Model { return admin.Catalog().Models() },
+							serve:   serve,
+							leader: func() bool {
+								return admin.Leadership() == nil || admin.Leadership().IsLeader()
+							},
+							gpus: func() []adminapi.ClassUse { return adminAPI.Summary().GPUs },
+							log:  log,
+						}
+						go keeper.run(ctx, 30*time.Second)
+					}
 					updateMetrics.Set(serve)
 					go serve.WatchUpdates(ctx, cfg.Mirror.UpdateCheck)
 				}
@@ -675,6 +747,14 @@ func newServerCmd() *cobra.Command {
 						return false
 					}
 					current.Store(&next)
+					if slices.ContainsFunc(changed, func(c string) bool { return strings.HasPrefix(c, "serving") }) {
+						rt.SetStopped(next.Serving.Stopped)
+						if admin != nil {
+							admin.Reconciler().SetStopped(next.Serving.Stopped)
+						}
+						log.Warn("the service is " + map[bool]string{true: "stopped", false: "started"}[next.Serving.Stopped] +
+							" by the cluster's settings")
+					}
 					if slices.ContainsFunc(changed, func(c string) bool { return strings.HasPrefix(c, "firewall") }) {
 						if err := writeFirewallState(next); err != nil {
 							log.Warn("could not record the firewall's settings for this machine", "err", err)
@@ -682,6 +762,13 @@ func newServerCmd() *cobra.Command {
 						if agentSvc != nil {
 							agentSvc.BroadcastFirewall()
 						}
+						if fwPub != nil {
+							fwPub.Kick()
+						}
+					}
+					if slices.ContainsFunc(changed, func(c string) bool { return strings.HasPrefix(c, "agents") }) && agentSvc != nil {
+						// A moment apart per node, so off this goroutine.
+						go agentSvc.BroadcastAgentSettings()
 					}
 					if slices.ContainsFunc(changed, func(c string) bool { return strings.HasPrefix(c, "guard") }) {
 						if gate == nil {
@@ -1153,3 +1240,40 @@ func federationID(ctx context.Context, st *storepkg.DB) (string, error) {
 	// more of a secret than needed.
 	return secret[:16], nil
 }
+
+// embedder has a model embed one text through the gateway itself, the way a
+// kept request is replayed: for the administrator's knowledge-base
+// endpoints, not for anybody's traffic.
+func embedder(replay func(ctx context.Context, path, model string, body []byte) (int, []byte, time.Duration, error)) func(context.Context, string, string) ([]float32, error) {
+	return func(ctx context.Context, model, text string) ([]float32, error) {
+		body, _ := json.Marshal(map[string]any{"model": model, "input": []string{text}})
+		status, out, _, err := replay(ctx, "/v1/embeddings", model, body)
+		if err != nil {
+			return nil, err
+		}
+		if status != http.StatusOK {
+			msg := strings.TrimSpace(string(out))
+			var e struct {
+				Error struct {
+					Message string `json:"message"`
+				} `json:"error"`
+			}
+			if json.Unmarshal(out, &e) == nil && e.Error.Message != "" {
+				msg = e.Error.Message
+			}
+			return nil, fmt.Errorf("%d: %s", status, msg)
+		}
+		var res struct {
+			Data []struct {
+				Embedding []float32 `json:"embedding"`
+			} `json:"data"`
+		}
+		if err := json.Unmarshal(out, &res); err != nil {
+			return nil, err
+		}
+		if len(res.Data) != 1 || len(res.Data[0].Embedding) == 0 {
+			return nil, fmt.Errorf("%s answered no embedding", model)
+		}
+		return res.Data[0].Embedding, nil
+	}
+}

internal/cli/serveradd.go

@@ -160,7 +160,7 @@ func (sa *serverAdd) firewall(ctx context.Context) error {
 		return err
 	}
 	// What applies the cluster's firewall setting there when it changes.
-	if err := installFirewallSync(ctx, sa.remote); err != nil {
+	if err := installFirewallSync(ctx, sa.remote, firewallStateFile); err != nil {
 		fmt.Fprintf(sa.out, "  !!    firewall: %v\n", err)
 	}
 	if !cfg.Firewall.Enabled {

internal/cli/serverinit.go

@@ -293,7 +293,7 @@ func (in *serverInit) configJudgesWith(model string) bool {
 		return false
 	}
 	for _, p := range cfg.Guard.Policies {
-		if p.Judge != nil && p.Judge.Model == model {
+		if slices.Contains(p.Judge.Judges(), model) {
 			return true
 		}
 	}
@@ -308,18 +308,21 @@ func guardSection(judge bool) string {
       # judge: {model: ` + guardModelName + `, input: block, output: block, ...}
 `
 	if judge {
-		j = `      # A judge model reads what patterns cannot: violence, weapons, abuse,
-      # self-harm. Two replicas run on CPUs with AVX2, about half a second
-      # for a short text; an answer is read every 600 characters as it
-      # streams. With fail: closed, a request the judge cannot read within
+		j = `      # A judge model reads what patterns cannot: violence, crime, sexual
+      # content, self-harm, jailbreaks. Two replicas run on CPUs, about two
+      # seconds for a short text; an answer is read every 600 characters as
+      # it streams. With fail: closed, a request no judge can read within
       # latency_budget is refused (503); open serves it, flagged, unjudged.
-      # Defamation (S5), specialized advice (S6), intellectual property (S8)
-      # and elections (S13) are left out: they refuse too much ordinary work.
+      # PII is the patterns' (above), and copyright, unethical acts and
+      # politically sensitive topics are left out: they refuse too much
+      # ordinary work. Controversial counts as safe; controversial: unsafe
+      # makes it count. With a GPU: llmgw ctl guard judge enable --gpu, then
+      # llmgw ctl guard judge use ` + defaultGPUJudge + ` ` + guardModelName + `.
       judge:
-        model: ` + guardModelName + `
+        models: [` + guardModelName + `]
         input: block
         output: block
-        categories: [S1, S2, S3, S4, S7, S9, S10, S11, S12]
+        categories: [Violent, Non-violent Illegal Acts, Sexual Content or Sexual Acts, Suicide & Self-Harm, Jailbreak]
         latency_budget: 5s
 `
 	}

internal/cli/serverinit_cluster.go

@@ -41,11 +41,13 @@ const (
 	starterSource  = "hf:Qwen/Qwen2.5-0.5B-Instruct-GGUF/qwen2.5-0.5b-instruct-q4_k_m.gguf"
 	starterVersion = "starter"
 
-	// The guardrails' judge: Meta's Llama Guard 3 1B, which reads a prompt
-	// or an answer and says whether it is unsafe and why. Q4_K_M is about
-	// 900 MB and judges a short text in about half a second on a CPU.
-	guardModelName    = "llama-guard-3-1b"
-	guardModelSource  = "hf:QuantFactory/Llama-Guard-3-1B-GGUF/Llama-Guard-3-1B.Q4_K_M.gguf"
+	// The guardrails' judge: Qwen3Guard-Gen 0.6B, which reads a prompt or an
+	// answer and says whether it is safe, controversial or unsafe, and why.
+	// Q8_0 is 0.8 GB and judges a short text in about two seconds on two CPU
+	// threads. It replaced Llama Guard 3 1B, which refused about one in two
+	// ordinary coding requests ("continue").
+	guardModelName    = defaultJudge
+	guardModelSource  = "hf:mradermacher/Qwen3Guard-Gen-0.6B-GGUF/Qwen3Guard-Gen-0.6B.Q8_0.gguf"
 	guardModelVersion = "guard"
 
 	// localLabel marks the control plane's own node, so the starter model is
@@ -443,12 +445,6 @@ func starterModel(source string, gpuClasses []string, pinned bool) catalog.Model
 	}
 }
 
-// llamaGuard3Categories name Llama Guard 3's S1 to S13, so a verdict is
-// recorded as "Violent Crimes" rather than "S1".
-var llamaGuard3Categories = []string{"Violent Crimes", "Non-Violent Crimes", "Sex-Related Crimes",
-	"Child Sexual Exploitation", "Defamation", "Specialized Advice", "Privacy", "Intellectual Property",
-	"Indiscriminate Weapons", "Hate", "Suicide & Self-Harm", "Sexual Content", "Elections"}
-
 // guardModel is the judge's catalog entry: CPU only, so it takes no card from
 // the models people use, on whichever node has room.
 func guardModel(source string) catalog.Model {
@@ -457,7 +453,7 @@ func guardModel(source string) catalog.Model {
 		Description: "The guardrails' judge: reads prompts and answers for unsafe content. Not offered to clients.",
 		Source:      source,
 		Role:        catalog.RoleGuard,
-		Guard:       &catalog.GuardModel{Format: "llamaguard3", Categories: llamaGuard3Categories},
+		Guard:       &catalog.GuardModel{Format: "qwen3guard"},
 		// Two: every judged request waits on it, and with fail: closed a
 		// judge that is not there refuses everything.
 		Scaling: catalog.Scaling{Min: 2, Max: 2},

internal/cli/serverinit_cluster_test.go

@@ -7,12 +7,14 @@ import (
 	"net/http/httptest"
 	"os"
 	"path/filepath"
+	"slices"
 	"strings"
 	"testing"
 	"time"
 
 	"gitea.mynodes.xyz/ilker/llmgw/internal/config"
 	"gitea.mynodes.xyz/ilker/llmgw/internal/control/catalog"
+	"gitea.mynodes.xyz/ilker/llmgw/internal/guard"
 )
 
 func TestTheInitAgentUnitIsTheShippedUnit(t *testing.T) {
@@ -150,9 +152,17 @@ func TestTheInstalledConfigurationIsHardenedAndLoads(t *testing.T) {
 		if (p.Judge != nil) != judge {
 			t.Errorf("judge=%v: the policy's judge is %+v", judge, p.Judge)
 		}
-		if judge && (p.Judge.Model != guardModelName || p.Judge.Input != "block") {
+		if judge && (!slices.Equal(p.Judge.Judges(), []string{guardModelName}) || p.Judge.Input != "block") {
 			t.Errorf("judge %+v", p.Judge)
 		}
+		// A category the judge never names would never count.
+		if judge {
+			for _, c := range p.Judge.Categories {
+				if !slices.Contains(guard.DefaultCategories("qwen3guard"), c) {
+					t.Errorf("the baseline counts %q, which Qwen3Guard does not name", c)
+				}
+			}
+		}
 		if cfg.Auth.MaxKeyTTL != 90*24*time.Hour || cfg.Auth.SignupOpen || cfg.DebugPprof || cfg.AdminRemote {
 			t.Errorf("not hardened: ttl %v signup %v pprof %v remote %v",
 				cfg.Auth.MaxKeyTTL, cfg.Auth.SignupOpen, cfg.DebugPprof, cfg.AdminRemote)
@@ -161,13 +171,13 @@ func TestTheInstalledConfigurationIsHardenedAndLoads(t *testing.T) {
 }
 
 func TestTheGuardModelIsAValidJudge(t *testing.T) {
-	m := guardModel("http://10.0.0.5:8081/llama-guard-3-1b/guard/g.gguf")
+	m := guardModel("http://10.0.0.5:8081/qwen3guard-0.6b/guard/g.gguf")
 	c, err := catalog.New([]catalog.Model{m})
 	if err != nil {
 		t.Fatalf("not a valid catalog entry: %v", err)
 	}
 	got, _ := c.Resolve(guardModelName)
-	if !got.IsGuard() || got.Guard.Format != "llamaguard3" || len(got.Guard.Categories) != 13 || got.Scaling.Min != 2 {
+	if !got.IsGuard() || got.Guard.Format != "qwen3guard" || got.Scaling.Min != 2 {
 		t.Errorf("%+v", got)
 	}
 	for _, p := range got.Profiles {

internal/cli/serverinit_firewall.go

@@ -93,6 +93,11 @@ func (in *serverInit) storeFirewall(ctx context.Context, members []pgha.Node, se
 				} else {
 					in.kept("firewall on %s", m.Name)
 				}
+				// And after, as the cluster's settings say, read from the
+				// member's own copy of the store.
+				if err := firewall.InstallFollower(ctx, m.Remote); err != nil {
+					fmt.Fprintf(in.out, "  !!    firewall on %s: %v; the cluster's settings will not change it\n", m.Name, err)
+				}
 				continue
 			}
 		}
@@ -135,5 +140,5 @@ func loopbackListen(listen string) bool {
 // ensureFirewallSync installs the units that apply the cluster's firewall
 // setting on this control plane.
 func (in *serverInit) ensureFirewallSync(ctx context.Context) error {
-	return installFirewallSync(ctx, firewall.Local{})
+	return installFirewallSync(ctx, firewall.Local{}, firewallStateFile)
 }

internal/cli/serverinit_postgres.go

@@ -16,6 +16,7 @@ import (
 	"gitea.mynodes.xyz/ilker/llmgw/internal/agent/discovery"
 	"gitea.mynodes.xyz/ilker/llmgw/internal/agent/provision"
 	"gitea.mynodes.xyz/ilker/llmgw/internal/config"
+	"gitea.mynodes.xyz/ilker/llmgw/internal/pgha"
 	"gitea.mynodes.xyz/ilker/llmgw/internal/store"
 )
 
@@ -348,15 +349,23 @@ func (in *serverInit) setupPostgres(ctx context.Context) (string, error) {
 	}
 
 	if installed {
-		// Its defaults assume it owns the machine. A gateway swapped out to
-		// give the database cache is worse than a smaller cache, and a store
-		// whose working set is megabytes needs little. Only a PostgreSQL init
-		// installed itself: one that was already here may serve something
-		// else that needs its connections.
-		for _, q := range []string{
-			"ALTER SYSTEM SET shared_buffers = '96MB'",
-			"ALTER SYSTEM SET max_connections = 40",
-		} {
+		// Sized to the machine's memory, as one it shares with the gateway:
+		// a gateway swapped out to give the database cache is worse than a
+		// smaller cache, but a knowledge base's vector index is searched
+		// well only from memory. Only a PostgreSQL init installed itself:
+		// one that was already here may serve something else that needs its
+		// connections.
+		meminfo, _ := os.ReadFile("/proc/meminfo")
+		total, ok := pgha.MemTotal(string(meminfo))
+		if !ok {
+			total = 1 << 30
+		}
+		mem := pgha.MemorySettings(total, true)
+		qs := []string{"ALTER SYSTEM SET max_connections = 40"}
+		for _, k := range []string{"shared_buffers", "effective_cache_size", "maintenance_work_mem", "work_mem"} {
+			qs = append(qs, "ALTER SYSTEM SET "+k+" = '"+mem[k]+"'")
+		}
+		for _, q := range qs {
 			if _, err := in.psql(ctx, q); err != nil {
 				return "", err
 			}
@@ -367,9 +376,12 @@ func (in *serverInit) setupPostgres(ctx context.Context) (string, error) {
 		if err := in.waitForPostgres(ctx); err != nil {
 			return "", err
 		}
-		in.done("PostgreSQL sized for a machine it shares (shared_buffers 96MB, 40 connections)")
+		in.done("PostgreSQL sized for a machine it shares, with %d MB of memory (shared_buffers %s, work_mem %s, 40 connections)",
+			total>>20, mem["shared_buffers"], mem["work_mem"])
 	}
 
+	in.setupPgvector(ctx, discovery.DiscoverHost())
+
 	// Proof the address works, before anything is moved into it.
 	address := "@" + in.dsnPath()
 	s, err := store.Open(ctx, address)
@@ -482,8 +494,17 @@ func lastLines(b []byte, n int) string {
 
 // psql runs one statement as the postgres superuser, over the local socket.
 func (in *serverInit) psql(ctx context.Context, query string) (string, error) {
-	cmd := exec.CommandContext(ctx, "runuser", "-u", "postgres", "--",
-		in.pg.tool("psql"), "-v", "ON_ERROR_STOP=1", "-qAtX", "-c", query)
+	return in.psqlIn(ctx, "", query)
+}
+
+// psqlIn runs a statement as postgres in one database, or in postgres's
+// default one when db is "".
+func (in *serverInit) psqlIn(ctx context.Context, db, query string) (string, error) {
+	args := []string{"-v", "ON_ERROR_STOP=1", "-qAtX", "-c", query}
+	if db != "" {
+		args = append(args, "-d", db)
+	}
+	cmd := exec.CommandContext(ctx, "runuser", append([]string{"-u", "postgres", "--", in.pg.tool("psql")}, args...)...)
 	// postgres cannot enter root's home, and psql says so on every call.
 	cmd.Dir = "/"
 	out, err := cmd.CombinedOutput()
@@ -538,3 +559,50 @@ func (in *serverInit) stopService(ctx context.Context) (bool, error) {
 	in.done("llmgw-server stopped while its store moves")
 	return true, nil
 }
+
+// pgvectorPackages are the package names of pgvector for a PostgreSQL major
+// version: Debian's and Ubuntu's (and PGDG's apt repository) name it after the
+// server, PGDG's rpm repository after the version alone.
+func pgvectorPackages(major int) map[string][]string {
+	return map[string][]string{
+		"deb": {fmt.Sprintf("postgresql-%d-pgvector", major)},
+		"rpm": {fmt.Sprintf("pgvector_%d", major)},
+	}
+}
+
+// setupPgvector installs pgvector and enables it in llmgw's database, for
+// knowledge bases. Not a reason to stop: without it everything else works,
+// and the store says why knowledge bases are off when somebody asks for one.
+func (in *serverInit) setupPgvector(ctx context.Context, h discovery.Host) {
+	warn := func(format string, a ...any) {
+		fmt.Fprintf(in.out, "  !!    pgvector: "+format+"; knowledge bases stay off\n", a...)
+	}
+	if on, err := in.psqlIn(ctx, pgDatabase, "SELECT 1 FROM pg_extension WHERE extname = 'vector'"); err == nil && on == "1" {
+		in.kept("pgvector in database %s", pgDatabase)
+		return
+	}
+	have, err := in.psqlBool(ctx, "SELECT 1 FROM pg_available_extensions WHERE name = 'vector'")
+	if err != nil {
+		warn("%v", err)
+		return
+	}
+	if !have {
+		num, err := in.psql(ctx, "SHOW server_version_num")
+		n, convErr := strconv.Atoi(num)
+		if err != nil || convErr != nil || h.Family() == "" {
+			warn("cannot tell which package to install (version %q, %s)", num, h.OSID)
+			return
+		}
+		pkgs := pgvectorPackages(n / 10000)
+		pi := &provision.Installer{Host: h, Log: slog.Default()}
+		if err := pi.InstallPackages(ctx, pkgs); err != nil {
+			warn("installing %s: %v", strings.Join(pkgs[h.Family()], " "), err)
+			return
+		}
+	}
+	if _, err := in.psqlIn(ctx, pgDatabase, "CREATE EXTENSION IF NOT EXISTS vector"); err != nil {
+		warn("%v", err)
+		return
+	}
+	in.done("pgvector enabled in database %s", pgDatabase)
+}

internal/cli/service.go

@@ -0,0 +1,187 @@
+package cli
+
+import (
+	"context"
+	"errors"
+	"fmt"
+	"os"
+	"os/exec"
+	"runtime"
+	"strings"
+
+	"github.com/spf13/cobra"
+)
+
+// llmgwServices are the units llmgw runs as on a machine, in the order they
+// start: the store's clients after the store, the server before the agent
+// that dials it.
+var llmgwServices = []string{"llmgw-mirror", "llmgw-server", "llmgw-agent"}
+
+// unitDir is where llmgw's units are installed.
+var unitDir = "/etc/systemd/system"
+
+// serviceRunner runs systemctl; a variable so tests stand in for it.
+var serviceRunner = func(ctx context.Context, args ...string) (string, error) {
+	out, err := exec.CommandContext(ctx, "systemctl", args...).CombinedOutput() //nolint:gosec // fixed command
+	return strings.TrimSpace(string(out)), err
+}
+
+// installedServices are the llmgw units this machine has, or those named.
+func installedServices(ctx context.Context, names []string) ([]string, error) {
+	var want []string
+	for _, n := range names {
+		if !strings.HasPrefix(n, "llmgw-") {
+			n = "llmgw-" + n
+		}
+		want = append(want, n)
+	}
+	var have []string
+	for _, u := range llmgwServices {
+		if _, err := os.Stat(unitDir + "/" + u + ".service"); err != nil {
+			continue
+		}
+		if len(want) > 0 && !hasUnit(want, u) {
+			continue
+		}
+		have = append(have, u)
+	}
+	for _, w := range want {
+		if !hasUnit(have, w) {
+			return nil, fmt.Errorf("%s is not installed here (this machine has: %s)", w, strings.Join(present(), ", "))
+		}
+	}
+	if len(have) == 0 {
+		return nil, errors.New("no llmgw service is installed on this machine")
+	}
+	return have, nil
+}
+
+func present() []string {
+	var out []string
+	for _, u := range llmgwServices {
+		if _, err := os.Stat(unitDir + "/" + u + ".service"); err == nil {
+			out = append(out, strings.TrimPrefix(u, "llmgw-"))
+		}
+	}
+	if len(out) == 0 {
+		return []string{"none"}
+	}
+	return out
+}
+
+func hasUnit(list []string, s string) bool {
+	for _, x := range list {
+		if x == s {
+			return true
+		}
+	}
+	return false
+}
+
+// newServiceCmd is llmgw service: this machine's llmgw services, one by one.
+// The whole system is llmgw status, start, stop and restart.
+func newServiceCmd() *cobra.Command {
+	c := &cobra.Command{
+		Use:   "service",
+		Short: "This machine's llmgw services: status, start, stop, restart",
+		Long: "The control plane (server), node agent (agent) and model mirror (mirror) on\n" +
+			"this machine alone, by systemd. The whole system is 'llmgw status', 'llmgw\n" +
+			"stop', 'llmgw start' and 'llmgw restart', run on a control plane.",
+	}
+	c.AddCommand(serviceVerbs()...)
+	return c
+}
+
+func serviceVerbs() []*cobra.Command {
+	status := &cobra.Command{
+		Use:   "status [server|agent|mirror]...",
+		Short: "Whether llmgw's services on this machine run",
+		Long: "Each llmgw service this machine has — the control plane (server), the\n" +
+			"node agent (agent), the model mirror (mirror) — and whether it runs.",
+		RunE: func(cmd *cobra.Command, args []string) error {
+			units, err := localUnits(cmd, args)
+			if err != nil {
+				return err
+			}
+			out := cmd.OutOrStdout()
+			for _, u := range units {
+				state, _ := serviceRunner(cmd.Context(), "is-active", u)
+				since, _ := serviceRunner(cmd.Context(), "show", u, "-p", "ActiveEnterTimestamp", "--value")
+				fmt.Fprintf(out, "%-14s %-10s %s\n", strings.TrimPrefix(u, "llmgw-"), state, since)
+			}
+			return nil
+		},
+	}
+	verb := func(use, short, long, action string, reverse bool) *cobra.Command {
+		var engines bool
+		c := &cobra.Command{
+			Use:   use + " [server|agent|mirror]...",
+			Short: short,
+			Long:  long,
+			RunE: func(cmd *cobra.Command, args []string) error {
+				if os.Geteuid() != 0 {
+					return errors.New("starting and stopping services needs root: run it with sudo")
+				}
+				units, err := localUnits(cmd, args)
+				if err != nil {
+					return err
+				}
+				if reverse {
+					// The agent before the server it dials, the server before
+					// the mirror it reads.
+					for i, j := 0, len(units)-1; i < j; i, j = i+1, j-1 {
+						units[i], units[j] = units[j], units[i]
+					}
+				}
+				out := cmd.OutOrStdout()
+				for _, u := range units {
+					if msg, err := serviceRunner(cmd.Context(), action, u); err != nil {
+						return fmt.Errorf("%s %s: %v %s", action, u, err, msg)
+					}
+					state, _ := serviceRunner(cmd.Context(), "is-active", u)
+					fmt.Fprintf(out, "%-14s %s\n", strings.TrimPrefix(u, "llmgw-"), state)
+				}
+				if action == "stop" && hasUnit(units, "llmgw-agent") {
+					if engines {
+						_, _ = exec.CommandContext(cmd.Context(), "pkill", "-u", "llmgw-engine").CombinedOutput() //nolint:gosec // fixed command
+						fmt.Fprintln(out, "engines   stopped (every process of llmgw-engine)")
+					} else {
+						fmt.Fprintln(out, "\nThe engines keep serving: they outlive the agent, which takes them over again when it starts.\n"+
+							"--engines stops them too; the control plane then places their replicas elsewhere.")
+					}
+				}
+				if action == "stop" && hasUnit(units, "llmgw-server") {
+					fmt.Fprintln(out, "A standby control plane, where there is one, takes over within seconds.")
+				}
+				return nil
+			},
+		}
+		if action == "stop" {
+			c.Flags().BoolVar(&engines, "engines", false, "also stop the engines this node runs")
+		}
+		return c
+	}
+	return []*cobra.Command{
+		status,
+		verb("start", "Start llmgw's services on this machine",
+			"Starts each llmgw service this machine has, or those named: the mirror, the\n"+
+				"control plane, the node agent, in that order.", "start", false),
+		verb("stop", "Stop llmgw's services on this machine",
+			"Stops each llmgw service this machine has, or those named: the agent, the\n"+
+				"control plane, the mirror, in that order. The engines a node runs keep\n"+
+				"serving unless --engines is given; a standby control plane takes over from\n"+
+				"a leader stopped. They stay stopped until 'llmgw start', or the machine\n"+
+				"boots.", "stop", true),
+		verb("restart", "Restart llmgw's services on this machine",
+			"Restarts each llmgw service this machine has, or those named. Engines keep\n"+
+				"serving and are taken over again by the agent; a control plane's settings\n"+
+				"and server.yaml are read anew.", "restart", false),
+	}
+}
+
+func localUnits(cmd *cobra.Command, args []string) ([]string, error) {
+	if runtime.GOOS != "linux" {
+		return nil, fmt.Errorf("llmgw's services run under systemd on Linux; this is %s", runtime.GOOS)
+	}
+	return installedServices(cmd.Context(), args)
+}

internal/cli/service_test.go

@@ -0,0 +1,137 @@
+package cli
+
+import (
+	"bytes"
+	"context"
+	"encoding/json"
+	"net/http"
+	"net/http/httptest"
+	"os"
+	"path/filepath"
+	"runtime"
+	"strings"
+	"testing"
+	"time"
+
+	"github.com/spf13/cobra"
+
+	"gitea.mynodes.xyz/ilker/llmgw/internal/control/adminapi"
+)
+
+func TestTheServicesHereAreFoundAndAskedAbout(t *testing.T) {
+	old, oldRun := unitDir, serviceRunner
+	unitDir = t.TempDir()
+	t.Cleanup(func() { unitDir, serviceRunner = old, oldRun })
+	for _, u := range []string{"llmgw-server", "llmgw-agent"} {
+		_ = os.WriteFile(filepath.Join(unitDir, u+".service"), []byte("[Unit]"), 0o644)
+	}
+	var asked []string
+	serviceRunner = func(_ context.Context, args ...string) (string, error) {
+		asked = append(asked, strings.Join(args, " "))
+		if args[0] == "is-active" {
+			return "active", nil
+		}
+		return "Wed 2026-10-07 12:00:00 UTC", nil
+	}
+
+	units, err := installedServices(context.Background(), nil)
+	if err != nil || strings.Join(units, ",") != "llmgw-server,llmgw-agent" {
+		t.Fatalf("%v %v", units, err)
+	}
+	if _, err := installedServices(context.Background(), []string{"mirror"}); err == nil ||
+		!strings.Contains(err.Error(), "server, agent") {
+		t.Errorf("a service not here: %v", err)
+	}
+	if u, err := installedServices(context.Background(), []string{"agent"}); err != nil || len(u) != 1 {
+		t.Errorf("one named: %v %v", u, err)
+	}
+
+	if runtime.GOOS != "linux" {
+		return
+	}
+	var out bytes.Buffer
+	c := serviceVerbs()[0]
+	c.SetOut(&out)
+	c.SetContext(context.Background())
+	if err := c.RunE(c, nil); err != nil {
+		t.Fatal(err)
+	}
+	if !strings.Contains(out.String(), "server") || !strings.Contains(out.String(), "active") {
+		t.Errorf("%s", out.String())
+	}
+}
+
+func TestStopSetsTheClustersServingStopped(t *testing.T) {
+	var put adminapi.SettingsRequest
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		switch {
+		case r.URL.Path == "/api/v1/summary":
+			_ = json.NewEncoder(w).Encode(adminapi.Summary{})
+		case r.Method == "GET" && r.URL.Path == "/api/v1/settings":
+			_ = json.NewEncoder(w).Encode(adminapi.SettingsView{Version: 3, Body: "router:\n  retries: 2\n"})
+		case r.Method == "PUT" && r.URL.Path == "/api/v1/settings":
+			_ = json.NewDecoder(r.Body).Decode(&put)
+			_ = json.NewEncoder(w).Encode(adminapi.SettingsView{Version: 4, Body: put.Body})
+		default:
+			http.NotFound(w, r)
+		}
+	}))
+	defer srv.Close()
+	dir := t.TempDir()
+	cfg := filepath.Join(dir, "server.yaml")
+	_ = os.WriteFile(filepath.Join(dir, "token"), []byte("t"), 0o600)
+	_ = os.WriteFile(cfg, []byte("listen: \":8080\"\nagent_insecure: true\nadmin_listen: "+
+		strings.TrimPrefix(srv.URL, "http://")+"\nadmin_remote: true\nadmin_token_file: "+filepath.Join(dir, "token")+"\n"), 0o600)
+
+	var stop *cobra.Command
+	for _, c := range newSystemCmds() {
+		if c.Name() == "stop" {
+			stop = c
+		}
+	}
+	var out bytes.Buffer
+	stop.SetOut(&out)
+	stop.SetArgs([]string{"--config", cfg, "--note", "maintenance"})
+	if err := stop.ExecuteContext(context.Background()); err != nil {
+		t.Fatalf("%v: %s", err, out.String())
+	}
+	if put.Base != 3 || !strings.Contains(put.Body, "stopped: true") || !strings.Contains(put.Body, "retries: 2") ||
+		put.Note != "maintenance" {
+		t.Errorf("sent %+v", put)
+	}
+	if !strings.Contains(out.String(), "stopping") {
+		t.Errorf("%s", out.String())
+	}
+}
+
+// Started while the control plane has not yet taken the setting up: every
+// model of a stopped system has the nothing it asks for, which is not
+// serving.
+func TestStartWaitsForTheControlPlaneToServe(t *testing.T) {
+	old := servicePoll
+	servicePoll = time.Millisecond
+	t.Cleanup(func() { servicePoll = old })
+	answers := []adminapi.Summary{
+		{Stopped: true, Models: adminapi.ModelTotals{Total: 2, Satisfied: 2}},
+		{Models: adminapi.ModelTotals{Total: 2, Satisfied: 0}},
+		{Models: adminapi.ModelTotals{Total: 2, Satisfied: 2}},
+	}
+	asked := 0
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
+		_ = json.NewEncoder(w).Encode(answers[min(asked, len(answers)-1)])
+		asked++
+	}))
+	defer srv.Close()
+
+	var out bytes.Buffer
+	cmd := &cobra.Command{}
+	cmd.SetOut(&out)
+	cmd.SetContext(context.Background())
+	c := &ctlClient{base: srv.URL, http: srv.Client()}
+	if err := waitForService(cmd, func() *ctlClient { return c }, false); err != nil {
+		t.Fatal(err)
+	}
+	if asked != 3 {
+		t.Errorf("returned after %d answers: %s", asked, out.String())
+	}
+}

internal/cli/status.go

@@ -153,6 +153,19 @@ func printSummary(cmd *cobra.Command, s adminapi.Summary) {
 		fmt.Fprintf(out, "STORE     %s\n", strings.Join(parts, ", "))
 	}
 
+	if s.Stopped {
+		fmt.Fprintln(out, "SERVICE   STOPPED: no model runs and requests are answered 503 (llmgw start)")
+	}
+	if s.GuardOff {
+		fmt.Fprintln(out, "GUARD     off: no request is checked, whatever the policies say (llmgw ctl guard on)")
+	}
+	if len(s.Judges) > 0 {
+		fmt.Fprintln(out, "\nGUARDRAILS' JUDGE")
+		for _, j := range s.Judges {
+			fmt.Fprintf(out, "  %s: %s\n", j.Model, j.Problem)
+		}
+	}
+
 	if len(s.Parked) > 0 {
 		fmt.Fprintln(out, "\nPARKED — failed every time they started; not retried until fixed")
 		for _, p := range s.Parked {

internal/cli/system.go

@@ -0,0 +1,232 @@
+package cli
+
+import (
+	"errors"
+	"fmt"
+	"os"
+	"os/exec"
+	"strings"
+	"time"
+
+	"github.com/spf13/cobra"
+
+	"gitea.mynodes.xyz/ilker/llmgw/internal/control/adminapi"
+	"gitea.mynodes.xyz/ilker/llmgw/internal/onboard"
+	"gitea.mynodes.xyz/ilker/llmgw/internal/settings"
+)
+
+// systemClient is the admin API of the control plane on this machine, as its
+// server.yaml says, with its token.
+func systemClient(configPath string) func() *ctlClient {
+	return func() *ctlClient {
+		addr := "127.0.0.1:9090"
+		var tokens []string
+		if cfg, ok := loadServerConfig(configPath); ok {
+			if cfg.AdminListen != "" {
+				addr = cfg.AdminListen
+				if strings.HasPrefix(addr, ":") {
+					addr = "127.0.0.1" + addr
+				}
+			}
+			tokens = append(tokens, adminTokenPath(configPath, cfg))
+		}
+		return newCtlClient(addr, tokens...)
+	}
+}
+
+// newSystemCmds are llmgw status, stop, start and restart: the whole system,
+// from a control plane.
+func newSystemCmds() []*cobra.Command {
+	var configPath string
+	client := systemClient("/etc/llmgw/server.yaml")
+	bind := func(c *cobra.Command) *cobra.Command {
+		c.Flags().StringVar(&configPath, "config", "/etc/llmgw/server.yaml", "this control plane's configuration")
+		c.PreRun = func(*cobra.Command, []string) { client = systemClient(configPath) }
+		return c
+	}
+	call := func() *ctlClient { return client() }
+
+	status := bind(&cobra.Command{
+		Use:   "status",
+		Short: "The whole system: whether it serves, its control planes, nodes, models and store",
+		Long: "Run on a control plane. Says whether the service is stopped (llmgw stop),\n" +
+			"and everything 'llmgw ctl status' does. One machine's services are\n" +
+			"'llmgw service status'.",
+		Args: cobra.NoArgs,
+		RunE: func(cmd *cobra.Command, _ []string) error {
+			var s adminapi.Summary
+			if err := call().get(cmd.Context(), "/api/v1/summary", &s); err != nil {
+				return notAControlPlane(err)
+			}
+			printSummary(cmd, s)
+			return nil
+		},
+	})
+
+	switchService := func(stop bool) *cobra.Command {
+		use, short, long := "start", "Start the whole system serving again, after llmgw stop",
+			"Every model comes back as its catalog entry asks: placed, loaded and\n"+
+				"serving again, on every control plane within seconds. Run on a control\n"+
+				"plane. The services themselves are never stopped by llmgw stop; a\n"+
+				"machine's own are started with 'llmgw service start'."
+		if stop {
+			use, short, long = "stop", "Stop the whole system serving: every model taken down, requests answered 503",
+				"Every model's replicas are taken down, freeing every GPU, and every\n"+
+					"request for a model is answered 503 (system_stopped) until 'llmgw start'.\n"+
+					"The control planes, agents and mirror keep running, so the console, the\n"+
+					"admin pages and llmgw start still work, and no model's definition is\n"+
+					"changed. In force on every control plane within seconds; kept in the\n"+
+					"cluster's settings (serving.stopped), so a restart does not undo it."
+		}
+		var note string
+		var wait bool
+		c := bind(&cobra.Command{
+			Use:   use,
+			Short: short,
+			Long:  long,
+			Args:  cobra.NoArgs,
+			RunE: func(cmd *cobra.Command, _ []string) error {
+				if note == "" {
+					note = "llmgw " + use
+				}
+				var s adminapi.Summary
+				if err := call().get(cmd.Context(), "/api/v1/summary", &s); err != nil {
+					return notAControlPlane(err)
+				}
+				if s.Stopped == stop {
+					fmt.Fprintf(cmd.OutOrStdout(), "the system is already %s\n", map[bool]string{true: "stopped", false: "serving"}[stop])
+					return nil
+				}
+				if err := changeSettings(cmd, call, note, func(doc *settings.Doc) error {
+					return doc.SetYAML([]string{"serving", "stopped"}, fmt.Sprint(stop))
+				}); err != nil {
+					return err
+				}
+				out := cmd.OutOrStdout()
+				if !wait {
+					if stop {
+						fmt.Fprintln(out, "stopping: every model's replicas are being taken down; llmgw status follows it")
+					} else {
+						fmt.Fprintln(out, "starting: every model is placed again as it asks; llmgw status follows it")
+					}
+					return nil
+				}
+				return waitForService(cmd, call, stop)
+			},
+		})
+		c.Flags().StringVar(&note, "note", "", "why, kept with the settings' version and in the audit")
+		c.Flags().BoolVar(&wait, "wait", false, "wait until every replica is down (stop) or the models have theirs (start)")
+		return c
+	}
+
+	var sshUser string
+	restart := bind(&cobra.Command{
+		Use:   "restart",
+		Short: "Restart every control plane and node agent; the engines keep serving",
+		Long: "Every node's agent starts again, then the other control planes over ssh\n" +
+			"as --ssh-user, then this one — the leader last, so it hands over to a\n" +
+			"standby already restarted. The engines keep serving throughout and are\n" +
+			"taken over again; nothing is placed anew. Run on a control plane, with\n" +
+			"sudo. One machine's services are 'llmgw service restart'.",
+		Args: cobra.NoArgs,
+		RunE: func(cmd *cobra.Command, _ []string) error {
+			if os.Geteuid() != 0 {
+				return errors.New("restarting this control plane needs root: run it with sudo")
+			}
+			ctx, out := cmd.Context(), cmd.OutOrStdout()
+			var s adminapi.Summary
+			if err := call().get(ctx, "/api/v1/summary", &s); err != nil {
+				return notAControlPlane(err)
+			}
+			var agents adminapi.AgentRestartView
+			if err := call().post(ctx, "/api/v1/agents/restart", []byte("{}"), &agents); err != nil {
+				return fmt.Errorf("restarting the agents: %w", err)
+			}
+			fmt.Fprintf(out, "agents     %d told to restart: %s\n", len(agents.Nodes), strings.Join(agents.Nodes, ", "))
+
+			// The other control planes, standbys before a leader.
+			var others []adminapi.ControlPlane
+			for _, cp := range s.ControlPlanes.Members {
+				if !cp.This {
+					others = append(others, cp)
+				}
+			}
+			// One that cannot be reached is said and left: the rest of the
+			// system still restarts, and the leader still goes last.
+			var failed []string
+			for pass := 0; pass < 2; pass++ {
+				for _, cp := range others {
+					leading := cp.State == "leading"
+					if leading != (pass == 1) {
+						continue
+					}
+					r := &onboard.SSH{Addr: cp.Address, User: sshUser}
+					if res, err := r.Run(ctx, "sudo systemctl restart llmgw-server", nil); err != nil {
+						fmt.Fprintf(out, "%-10s NOT restarted: ssh as %q: %v %s\n", cp.Name, sshUser, err, strings.TrimSpace(string(res)))
+						failed = append(failed, cp.Name+" ("+cp.Address+")")
+						continue
+					}
+					fmt.Fprintf(out, "%-10s restarted (%s)\n", cp.Name, cp.State)
+					time.Sleep(5 * time.Second)
+				}
+			}
+			if _, err := os.Stat(unitDir + "/llmgw-server.service"); err == nil {
+				fmt.Fprintln(out, "this       restarting")
+				if b, err := exec.CommandContext(ctx, "systemctl", "restart", "llmgw-server").CombinedOutput(); err != nil { //nolint:gosec // fixed command
+					return fmt.Errorf("restarting this control plane: %v: %s", err, strings.TrimSpace(string(b)))
+				}
+				fmt.Fprintln(out, "this       restarted")
+			}
+			fmt.Fprintln(out, "\nllmgw status shows the nodes coming back; their replicas are taken over as they are.")
+			if len(failed) > 0 {
+				return fmt.Errorf("not restarted: %s; run `sudo llmgw service restart server` there, "+
+					"or give --ssh-user someone this machine's root can log in as", strings.Join(failed, ", "))
+			}
+			return nil
+		},
+	})
+	restart.Flags().StringVar(&sshUser, "ssh-user", os.Getenv("SUDO_USER"), "who to log in to the other control planes as")
+
+	return []*cobra.Command{status, switchService(true), switchService(false), restart}
+}
+
+// servicePoll is how often --wait asks; a variable so tests need not wait.
+var servicePoll = 5 * time.Second
+
+// waitForService waits until no replica runs (stopped) or every model has
+// the replicas it asks for (started).
+//
+// Only once the control plane says it is in the state asked for: it takes the
+// setting up at its next look at the store, seconds later, and until then a
+// stopped system's models, which ask for nothing, all have what they ask for —
+// so llmgw start --wait returned at once, with nothing started.
+func waitForService(cmd *cobra.Command, call func() *ctlClient, stop bool) error {
+	deadline := time.Now().Add(15 * time.Minute)
+	for time.Now().Before(deadline) {
+		var s adminapi.Summary
+		if err := call().get(cmd.Context(), "/api/v1/summary", &s); err == nil && s.Stopped == stop {
+			if stop && s.Replicas.Ready+s.Replicas.Starting+s.Replicas.Draining == 0 {
+				fmt.Fprintln(cmd.OutOrStdout(), "stopped: no replica runs")
+				return nil
+			}
+			if !stop && s.Models.Satisfied >= s.Models.Total {
+				fmt.Fprintf(cmd.OutOrStdout(), "serving: %d of %d models have the replicas they ask for\n",
+					s.Models.Satisfied, s.Models.Total)
+				return nil
+			}
+		}
+		select {
+		case <-cmd.Context().Done():
+			return cmd.Context().Err()
+		case <-time.After(servicePoll):
+		}
+	}
+	return errors.New("not there within fifteen minutes; llmgw status says where it stands")
+}
+
+// notAControlPlane is a failure to reach the admin API, said as what it
+// most often means.
+func notAControlPlane(err error) error {
+	return fmt.Errorf("%w\n(the whole system is stopped and started from a control plane; this machine's own "+
+		"services are 'llmgw service status|start|stop|restart')", err)
+}

internal/cli/toolcheck.go

@@ -34,10 +34,11 @@ import (
 // wrapper — left it in the content. A differently trained model fixes that. A
 // bigger one from the same family does not: qwen2.5-coder-7b invented
 // <function-call> and failed the same way.
-func newToolCheckCmd() *cobra.Command {
+func newToolCheckCmd(client func() *ctlClient) *cobra.Command {
 	var (
 		endpoint string
 		model    string
+		key      string
 		timeout  time.Duration
 	)
 
@@ -57,10 +58,21 @@ func newToolCheckCmd() *cobra.Command {
 			if endpoint == "" {
 				return fmt.Errorf("--endpoint is required, e.g. http://10.0.0.145:20002")
 			}
+			// A replica's engine demands its own key, and refused the check
+			// with a 401 when none was sent. Asked of the admin API, as bench
+			// does; an endpoint that is no replica's is asked without one.
+			if key == "" {
+				reps, err := listReplicas(cmd, client)
+				if err != nil {
+					cmd.PrintErrf("The admin API could not be asked for the engine key of the replica at "+
+						"--endpoint (%v); an engine that demands one will refuse. Pass --key to give it.\n", err)
+				}
+				key = replicaKey(reps, endpoint)
+			}
 			ctx, cancel := context.WithTimeout(cmd.Context(), timeout)
 			defer cancel()
 
-			res, err := probeToolCall(ctx, endpoint, model)
+			res, err := probeToolCall(ctx, endpoint, model, key)
 			if err != nil {
 				return err
 			}
@@ -75,6 +87,7 @@ func newToolCheckCmd() *cobra.Command {
 	}
 	cmd.Flags().StringVar(&endpoint, "endpoint", "", "engine base URL, e.g. http://10.0.0.145:20002")
 	cmd.Flags().StringVar(&model, "model", "", "model name to send, when the engine needs one")
+	cmd.Flags().StringVar(&key, "key", "", "the engine's key (default: the admin API's, for the replica at --endpoint)")
 	cmd.Flags().DurationVar(&timeout, "timeout", 5*time.Minute, "how long to wait; CPU inference is slow")
 	return cmd
 }
@@ -102,7 +115,7 @@ type ToolProbe struct {
 	Raw     string
 }
 
-func probeToolCall(ctx context.Context, endpoint, model string) (ToolProbe, error) {
+func probeToolCall(ctx context.Context, endpoint, model, key string) (ToolProbe, error) {
 	// A question with one obvious tool answer and no other sensible reply, so a
 	// model that can use tools has no excuse not to.
 	body := map[string]any{
@@ -138,6 +151,9 @@ func probeToolCall(ctx context.Context, endpoint, model string) (ToolProbe, erro
 		return ToolProbe{}, err
 	}
 	req.Header.Set("Content-Type", "application/json")
+	if key != "" {
+		req.Header.Set("Authorization", "Bearer "+key)
+	}
 
 	resp, err := http.DefaultClient.Do(req)
 	if err != nil {
@@ -150,6 +166,10 @@ func probeToolCall(ctx context.Context, endpoint, model string) (ToolProbe, erro
 		return ToolProbe{}, err
 	}
 	if resp.StatusCode != http.StatusOK {
+		if resp.StatusCode == http.StatusUnauthorized {
+			return ToolProbe{}, fmt.Errorf("%s answered 401: the engine's key is the admin API's, "+
+				"for the replica `llmgw ctl replicas` lists at --endpoint; or pass it with --key", url)
+		}
 		return ToolProbe{}, fmt.Errorf("%s returned %d: %s", url, resp.StatusCode, trim(string(raw), 300))
 	}
 	return readToolProbe(raw)

internal/cli/toolcheck_test.go

@@ -1,10 +1,65 @@
 package cli
 
 import (
+	"bytes"
+	"context"
+	"encoding/json"
+	"fmt"
+	"net/http"
+	"net/http/httptest"
 	"strings"
 	"testing"
+
+	"gitea.mynodes.xyz/ilker/llmgw/internal/control/adminapi"
 )
 
+// toolCheck runs the command against an engine that answers only its own key,
+// with an admin API that lists the replica there with that key.
+func toolCheck(t *testing.T, args ...string) (string, error) {
+	t.Helper()
+	engine := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		if r.Header.Get("Authorization") != "Bearer engine-key" {
+			http.Error(w, `{"error":{"message":"Invalid API Key"}}`, http.StatusUnauthorized)
+			return
+		}
+		fmt.Fprint(w, `{"choices":[{"message":{"tool_calls":[{"id":"c1","type":"function",
+		  "function":{"name":"get_weather","arguments":"{\"city\":\"Istanbul\"}"}}]}}]}`)
+	}))
+	admin := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		_ = json.NewEncoder(w).Encode([]adminapi.ReplicaView{
+			{ID: "r1", Model: "m", Profile: "gpu", Endpoint: engine.URL, EngineKey: "engine-key"},
+		})
+	}))
+	t.Cleanup(func() { engine.Close(); admin.Close() })
+	t.Setenv("LLMGW_ADMIN_TOKEN", "t")
+	cmd := newToolCheckCmd(func() *ctlClient { return newCtlClient(admin.URL) })
+	var out bytes.Buffer
+	cmd.SetOut(&out)
+	cmd.SetErr(&out)
+	cmd.SetArgs(append([]string{"--endpoint", engine.URL}, args...))
+	err := cmd.ExecuteContext(context.Background())
+	return out.String(), err
+}
+
+func TestToolCheckAsksWithTheEngineKeyTheAdminAPIGives(t *testing.T) {
+	// Every replica's engine demands its key, and the check was sent without
+	// one: it failed with a 401 on every engine in the cluster.
+	out, err := toolCheck(t)
+	if err != nil {
+		t.Fatalf("%v\n%s", err, out)
+	}
+	if !strings.Contains(out, "tool calls work") {
+		t.Errorf("unexpected report:\n%s", out)
+	}
+}
+
+func TestToolCheckSaysWhereTheEngineKeyComesFrom(t *testing.T) {
+	out, err := toolCheck(t, "--key", "wrong")
+	if err == nil || !strings.Contains(err.Error(), "401") || !strings.Contains(err.Error(), "--key") {
+		t.Fatalf("expected the refusal explained, got %v\n%s", err, out)
+	}
+}
+
 func TestAStructuredToolCallIsRecognised(t *testing.T) {
 	p, err := readToolProbe([]byte(`{"choices":[{"message":{"role":"assistant",
 	  "tool_calls":[{"id":"c1","type":"function","function":{"name":"get_weather","arguments":"{\"city\":\"Istanbul\"}"}}]}}]}`))

internal/cli/train.go

@@ -0,0 +1,293 @@
+package cli
+
+import (
+	"encoding/json"
+	"fmt"
+	"io"
+	"os"
+	"strings"
+	"time"
+
+	"github.com/spf13/cobra"
+
+	"gitea.mynodes.xyz/ilker/llmgw/internal/control/adminapi"
+	"gitea.mynodes.xyz/ilker/llmgw/internal/store"
+)
+
+func newTrainCmd(client func() *ctlClient, asJSON *bool) *cobra.Command {
+	c := &cobra.Command{
+		Use:   "train",
+		Short: "Train LoRA adapters on the nodes, on question/answer pairs",
+		Long: "A node trains an adapter for a catalog model on pairs (llmgw ctl kb dataset\n" +
+			"writes them from a knowledge base; read them through first). It trains with\n" +
+			"its installed vLLM's torch and a small overlay, as the engines' user, on a card\n" +
+			"with --gpu-memory free — beside an engine, if it leaves that much. What it\n" +
+			"makes is kept by the control plane: llmgw ctl train download, or publish it\n" +
+			"as the model's adapter.\n\n" +
+			"An adapter learns how answers sound — words, format, tone — far better than\n" +
+			"what they say. Facts belong to the knowledge base.",
+	}
+
+	var model, base, rev, node, data, owner string
+	var public bool
+	var gpuMem int
+	var epochs, rank int
+	var lr float64
+	var noGGUF bool
+	start := &cobra.Command{
+		Use:   "start NAME --model MODEL --base-repo OWNER/NAME --node NODE --data FILE",
+		Short: "Queue an adapter's training on a node",
+		Example: "  llmgw ctl train start policies-tone --model qwen2.5-0.5b-instruct \\\n" +
+			"    --base-repo Qwen/Qwen2.5-0.5B-Instruct --node cuda01 --gpu-memory 2300 --data policies.jsonl",
+		Args: cobra.ExactArgs(1),
+		RunE: func(cmd *cobra.Command, args []string) error {
+			b, err := os.ReadFile(data)
+			if err != nil {
+				return err
+			}
+			req := adminapi.TrainRequest{Name: args[0], Model: model, BaseRepo: base, BaseRevision: rev, Node: node,
+				GPUMemoryMiB: gpuMem, Data: string(b), Params: map[string]any{}, Owner: owner, Public: public}
+			if cmd.Flags().Changed("epochs") {
+				req.Params["epochs"] = epochs
+			}
+			if cmd.Flags().Changed("rank") {
+				req.Params["rank"] = rank
+			}
+			if cmd.Flags().Changed("lr") {
+				req.Params["lr"] = lr
+			}
+			if noGGUF {
+				f := false
+				req.GGUF = &f
+			}
+			body, _ := json.Marshal(req)
+			var v adminapi.TrainJobView
+			if err := client().patient(2*time.Minute).do(cmd.Context(), "POST", "/api/v1/train/jobs", json.RawMessage(body), &v); err != nil {
+				return err
+			}
+			who := "every caller of " + v.Model
+			if v.Owner != "" {
+				who = v.Owner + " only"
+			}
+			fmt.Fprintf(cmd.OutOrStdout(), "queued %s: %s for %s on %s, %d pairs; it will be seen by %s\n",
+				v.ID, v.Name, v.Model, v.Node, v.Pairs, who)
+			fmt.Fprintf(cmd.OutOrStdout(), "follow it with: llmgw ctl train show %s --follow\n", v.ID)
+			return nil
+		},
+	}
+	f := start.Flags()
+	f.StringVar(&model, "model", "", "the catalog model the adapter is for")
+	f.StringVar(&base, "base-repo", "", "the weights to train on, as safetensors on Hugging Face (owner/name)")
+	f.StringVar(&rev, "base-revision", "", "a revision of them (default: the newest)")
+	f.StringVar(&node, "node", "", "the node that trains")
+	f.IntVar(&gpuMem, "gpu-memory", 4096, "MiB of a card the job needs free")
+	f.StringVar(&data, "data", "", "the pairs, NDJSON of {\"q\", \"a\"}")
+	f.IntVar(&epochs, "epochs", 6, "passes over the pairs")
+	f.IntVar(&rank, "rank", 16, "the adapter's rank")
+	f.Float64Var(&lr, "lr", 2e-4, "the learning rate")
+	f.BoolVar(&noGGUF, "no-gguf", false, "skip the llama.cpp GGUF, even for a model llama.cpp serves")
+	f.StringVar(&owner, "owner", "", "who the adapter belongs to — org:, team:, project: (path or ID), user: or key: — and only they see it "+
+		"(default: the knowledge base's owner, when the pairs come from one)")
+	f.BoolVar(&public, "public", false, "every caller of the model sees it; refused for pairs from a knowledge base")
+	for _, req := range []string{"model", "base-repo", "node", "data"} {
+		_ = start.MarkFlagRequired(req)
+	}
+	c.AddCommand(start)
+
+	c.AddCommand(&cobra.Command{
+		Use:   "list",
+		Short: "Training jobs, newest first",
+		Args:  cobra.NoArgs,
+		RunE: func(cmd *cobra.Command, _ []string) error {
+			var jobs []adminapi.TrainJobView
+			if err := client().get(cmd.Context(), "/api/v1/train/jobs", &jobs); err != nil {
+				return err
+			}
+			if *asJSON {
+				return json.NewEncoder(cmd.OutOrStdout()).Encode(jobs)
+			}
+			if len(jobs) == 0 {
+				fmt.Fprintln(cmd.OutOrStdout(), "No training jobs (llmgw ctl train start).")
+				return nil
+			}
+			w := tab()
+			fmt.Fprintln(w, "JOB\tADAPTER\tMODEL\tOWNER\tNODE\tSTATE\tPROGRESS\tLOSS\tPEAK\tCREATED")
+			for _, j := range jobs {
+				fmt.Fprintf(w, "%s\t%s\t%s\t%s\t%s\t%s\t%s\t%s\t%s\t%s\n", j.ID, j.Name, j.Model, or(j.Owner, "everyone"), j.Node, j.State,
+					trainProgress(j), trainLoss(j), trainPeak(j), j.CreatedAt.Local().Format("01-02 15:04"))
+			}
+			return w.Flush()
+		},
+	})
+
+	var follow bool
+	show := &cobra.Command{
+		Use:   "show JOB",
+		Short: "How a training job goes, and what it made",
+		Args:  cobra.ExactArgs(1),
+		RunE: func(cmd *cobra.Command, args []string) error {
+			out := cmd.OutOrStdout()
+			last := ""
+			for {
+				var j adminapi.TrainJobView
+				if err := client().get(cmd.Context(), "/api/v1/train/jobs/"+args[0], &j); err != nil {
+					return err
+				}
+				if *asJSON {
+					return json.NewEncoder(out).Encode(j)
+				}
+				line := fmt.Sprintf("%s  %s  %s  loss %s  peak %s  %s", j.State, trainProgress(j), j.Node, trainLoss(j), trainPeak(j), j.Message)
+				if !follow {
+					fmt.Fprintf(out, "%s: adapter %s for %s, owned by %s, on %s (base %s), %d pairs\n", j.ID, j.Name, j.Model,
+						or(j.Owner, "everyone"), j.Node, j.BaseRepo, j.Pairs)
+					fmt.Fprintln(out, line)
+					if len(j.Artifacts) > 0 {
+						fmt.Fprintf(out, "made: %s (llmgw ctl train download %s --kind KIND)\n", strings.Join(j.Artifacts, ", "), j.ID)
+					}
+					return nil
+				}
+				if line != last {
+					fmt.Fprintf(out, "%s  %s\n", time.Now().Format("15:04:05"), line)
+					last = line
+				}
+				if store.TrainOver(j.State) {
+					if j.State != store.TrainDone {
+						return fmt.Errorf("%s %s", j.ID, j.State)
+					}
+					return nil
+				}
+				select {
+				case <-cmd.Context().Done():
+					return nil
+				case <-time.After(3 * time.Second):
+				}
+			}
+		},
+	}
+	show.Flags().BoolVarP(&follow, "follow", "f", false, "follow it until it ends")
+	c.AddCommand(show)
+
+	c.AddCommand(&cobra.Command{
+		Use:   "cancel JOB",
+		Short: "Stop a training job",
+		Args:  cobra.ExactArgs(1),
+		RunE: func(cmd *cobra.Command, args []string) error {
+			if err := client().do(cmd.Context(), "POST", "/api/v1/train/jobs/"+args[0]+"/cancel", nil, nil); err != nil {
+				return err
+			}
+			fmt.Fprintf(cmd.OutOrStdout(), "%s cancelled\n", args[0])
+			return nil
+		},
+	})
+	c.AddCommand(&cobra.Command{
+		Use:   "rm JOB",
+		Short: "Remove an ended training job and what it made",
+		Args:  cobra.ExactArgs(1),
+		RunE: func(cmd *cobra.Command, args []string) error {
+			if err := client().do(cmd.Context(), "DELETE", "/api/v1/train/jobs/"+args[0], nil, nil); err != nil {
+				return err
+			}
+			fmt.Fprintf(cmd.OutOrStdout(), "%s removed\n", args[0])
+			return nil
+		},
+	})
+
+	var restart bool
+	publish := &cobra.Command{
+		Use:   "publish JOB",
+		Short: "Make what a finished job made its model's adapter, kept on the mirror",
+		Long: "The GGUF for a model llama.cpp serves, the PEFT files for one vLLM serves, are kept\n" +
+			"on the mirror and added to the model's adapters; a request then asks for the\n" +
+			"adapter by its name, as a model. llama.cpp loads adapters when a replica starts:\n" +
+			"--restart replaces the model's replicas, one at a time, to load it.",
+		Args: cobra.ExactArgs(1),
+		RunE: func(cmd *cobra.Command, args []string) error {
+			var v adminapi.TrainPublishView
+			if err := client().patient(15*time.Minute).do(cmd.Context(), "POST", "/api/v1/train/jobs/"+args[0]+"/publish", nil, &v); err != nil {
+				return err
+			}
+			out := cmd.OutOrStdout()
+			fmt.Fprintf(out, "%s is %s's adapter, from %s\n", v.Adapter, v.Model, v.Source)
+			if v.Owner != "" {
+				fmt.Fprintf(out, "only %s sees it; to anyone else it is a model that does not exist\n", v.Owner)
+			}
+			if restart && strings.Contains(v.Note, "restart") {
+				body, _ := json.Marshal(adminapi.RestartRequest{Models: []string{v.Model}})
+				var res adminapi.RestartResult
+				if err := client().post(cmd.Context(), "/api/v1/restart", body, &res); err != nil {
+					return err
+				}
+				fmt.Fprintf(out, "%d replica(s) of %s restarting to load it; then ask for model %s\n", len(res.Queued), v.Model, v.Adapter)
+				return nil
+			}
+			fmt.Fprintln(out, v.Note)
+			return nil
+		},
+	}
+	publish.Flags().BoolVar(&restart, "restart", false, "restart the model's replicas to load it (llama.cpp)")
+	c.AddCommand(publish)
+
+	var kind, outFile string
+	download := &cobra.Command{
+		Use:   "download JOB",
+		Short: "Save what a training job made: the PEFT adapter (peft) or the GGUF (gguf)",
+		Args:  cobra.ExactArgs(1),
+		RunE: func(cmd *cobra.Command, args []string) error {
+			if outFile == "" {
+				outFile = args[0] + ".tar.gz"
+				if kind == "gguf" {
+					outFile = args[0] + ".gguf"
+				}
+			}
+			f, err := os.Create(outFile)
+			if err != nil {
+				return err
+			}
+			defer f.Close()
+			cw := &countWriter{w: f}
+			if err := client().patient(10*time.Minute).do(cmd.Context(), "GET",
+				"/api/v1/train/jobs/"+args[0]+"/artifacts/"+kind, nil, io.Writer(cw)); err != nil {
+				_ = os.Remove(outFile)
+				return err
+			}
+			fmt.Fprintf(cmd.OutOrStdout(), "%s: %s\n", outFile, humanBytes(cw.n))
+			return nil
+		},
+	}
+	download.Flags().StringVar(&kind, "kind", "peft", "peft (a gzipped tar of the PEFT adapter) or gguf")
+	download.Flags().StringVarP(&outFile, "out", "o", "", "the file to write")
+	c.AddCommand(download)
+	return c
+}
+
+type countWriter struct {
+	w io.Writer
+	n int64
+}
+
+func (c *countWriter) Write(p []byte) (int, error) {
+	n, err := c.w.Write(p)
+	c.n += int64(n)
+	return n, err
+}
+
+func trainProgress(j adminapi.TrainJobView) string {
+	if j.Steps == 0 {
+		return "-"
+	}
+	return fmt.Sprintf("%d/%d", j.Step, j.Steps)
+}
+
+func trainLoss(j adminapi.TrainJobView) string {
+	if j.Loss == 0 {
+		return "-"
+	}
+	return fmt.Sprintf("%.3f", j.Loss)
+}
+
+func trainPeak(j adminapi.TrainJobView) string {
+	if j.PeakMiB == 0 {
+		return "-"
+	}
+	return fmt.Sprintf("%d MiB", j.PeakMiB)
+}

internal/cli/tui_actions.go

@@ -679,6 +679,8 @@ func (m *consoleModel) doEntry() tea.Cmd {
 			}}
 	case cBackup:
 		return m.runAfter("server", "backup", "--config="+m.file)
+	case cJudge:
+		return m.runAfter("ctl", server, "guard", "judge", "enable", "--replace")
 	}
 	return nil
 }

internal/cli/tui_console.go

@@ -45,6 +45,7 @@ const (
 	cUsage
 	cQuotas
 	cGuard
+	cKB
 	cFederation
 	cMirror
 	cJobs
@@ -59,6 +60,7 @@ const (
 	cAddControl
 	cReplicate
 	cServe
+	cJudge
 	cUpgrade
 	cRestartAll
 	cBackup
@@ -67,8 +69,8 @@ const (
 	cCount
 )
 
-var consoleLabels = [cCount]string{"Overview", "Monitor", "Nodes", "GPUs", "Models", "Replicas", "Usage", "Quotas", "Guardrails", "Federation", "Mirror", "Jobs", "Logs", "Accounts", "API keys", "Organizations", "Audit", "Manual",
-	"server.yaml", "Add a node", "Add a control plane", "Replicate the store", "Serve a model", "Upgrade the cluster",
+var consoleLabels = [cCount]string{"Overview", "Monitor", "Nodes", "GPUs", "Models", "Replicas", "Usage", "Quotas", "Guardrails", "Knowledge", "Federation", "Mirror", "Jobs", "Logs", "Accounts", "API keys", "Organizations", "Audit", "Manual",
+	"server.yaml", "Add a node", "Add a control plane", "Replicate the store", "Serve a model", "Serve the guardrails' judge", "Upgrade the cluster",
 	"Restart all replicas", "Back up now", "Wipe this machine", "Leave"}
 
 // consoleAction is what is done once the screen closes: each runs its own
@@ -176,6 +178,7 @@ type consoleModel struct {
 	quotas    []adminapi.QuotaView
 	quotasErr error
 	guard     guardState
+	kb        kbState
 
 	st     *storeHolder
 	people peopleData
@@ -322,6 +325,9 @@ func (m consoleModel) Update(msg tea.Msg) (tea.Model, tea.Cmd) {
 		if m.menu == cGuard {
 			cmds = append(cmds, m.fetchGuard())
 		}
+		if m.menu == cKB {
+			cmds = append(cmds, m.fetchKB())
+		}
 		if m.menu == cMirror {
 			cmds = append(cmds, m.fetchMirror())
 		}
@@ -380,6 +386,8 @@ func (m consoleModel) Update(msg tea.Msg) (tea.Model, tea.Cmd) {
 		m.people = peopleData(msg)
 	case consoleQuotasMsg:
 		m.quotas, m.quotasErr = msg.rows, msg.err
+	case consoleKBMsg:
+		m.kb = kbState{view: msg.view, err: msg.err, loaded: true}
 	case consoleGuardMsg:
 		window := m.guard.window
 		m.guard = guardState{view: msg.view, events: msg.events, err: msg.err, window: window}
@@ -496,7 +504,7 @@ func (m consoleModel) key(k tea.KeyMsg) (tea.Model, tea.Cmd) {
 				return m.leave()
 			case cWipe:
 				return m.askWipe()
-			case cUpgrade, cRestartAll, cBackup:
+			case cUpgrade, cRestartAll, cBackup, cJudge:
 				return m, m.doEntry()
 			case cOverview:
 				return m, nil
@@ -538,6 +546,8 @@ func (m consoleModel) key(k tea.KeyMsg) (tea.Model, tea.Cmd) {
 			m.menuD = m.quotaMenu(m.quotas[m.row])
 		case cGuard:
 			m.menuD = m.guardAttachmentMenu(m.guard.view.Attachments[m.row])
+		case cKB:
+			m.menuD = m.kbMenu(m.kb.view.Collections[m.row])
 		case cMirror:
 			m.menuD = m.mirrorMenu(m.mirrorGroups()[m.row])
 		case cJobs:
@@ -595,6 +605,10 @@ func (m consoleModel) key(k tea.KeyMsg) (tea.Model, tea.Cmd) {
 			m.quotaForm(adminapi.QuotaView{})
 		case cGuard:
 			m.guardAttachForm()
+		case cKB:
+			if m.kb.view.Available {
+				m.kbCreateForm()
+			}
 		case cModels:
 			m.addModelForm()
 		case cMirror:
@@ -650,6 +664,8 @@ func (m consoleModel) onEnter() tea.Cmd {
 		return m.fetchQuotas()
 	case cGuard:
 		return m.fetchGuard()
+	case cKB:
+		return m.fetchKB()
 	case cMirror:
 		return m.fetchMirror()
 	case cMonitor:
@@ -1026,6 +1042,8 @@ func (m consoleModel) rowCount() int {
 		return len(m.quotas)
 	case cGuard:
 		return len(m.guard.view.Attachments)
+	case cKB:
+		return len(m.kb.view.Collections)
 	case cMirror:
 		return len(m.mirrorGroups())
 	case cJobs:
@@ -1229,13 +1247,17 @@ func (m consoleModel) panel() (title, body, keys string) {
 			for _, s := range r.Shards {
 				nodes = append(nodes, fmt.Sprintf("%s%v", s.Node, discovery.DeviceLabels(s.GPUs)))
 			}
-			rows = append(rows, []string{r.ID, r.Model, r.State, r.Pool, strings.Join(nodes, " "),
+			compute := "-"
+			if len(r.Shards) > 0 {
+				compute = r.Shards[0].Compute.Short()
+			}
+			rows = append(rows, []string{r.ID, r.Model, r.State, r.Pool, compute, strings.Join(nodes, " "),
 				(time.Duration(r.AgeSec) * time.Second).Round(time.Second).String()})
 		}
 		if m.inPanel {
 			keys = rowKeys
 		}
-		return "Replicas", "\n" + table([]string{"REPLICA", "MODEL", "STATE", "POOL", "ON", "AGE"}, rows, m.row, pw, m.inPanel), keys
+		return "Replicas", "\n" + table([]string{"REPLICA", "MODEL", "STATE", "POOL", "COMPUTE", "ON", "AGE"}, rows, m.row, pw, m.inPanel), keys
 	case cUsage:
 		head := "\n  Last 24 hours, by " + lipgloss.NewStyle().Bold(true).Render(usageGroupings[m.usageBy]) +
 			sDim.Render("   (→ to group by "+usageGroupings[(m.usageBy+1)%len(usageGroupings)]+")") + "\n\n"
@@ -1255,6 +1277,8 @@ func (m consoleModel) panel() (title, body, keys string) {
 		return m.quotasPanel(pw)
 	case cGuard:
 		return m.guardPanel(pw)
+	case cKB:
+		return m.kbPanel(pw)
 	case cAccounts, cKeys, cOrgs:
 		return m.peoplePanel(pw)
 	case cManual:
@@ -1299,6 +1323,9 @@ func (m consoleModel) panel() (title, body, keys string) {
 		item = "\n  Moves the store onto two machines under Patroni, replicated,\n  so losing one loses nothing.\n\n  " +
 			sDim.Render("On the command line:") +
 			"\n  sudo --preserve-env=SSH_AUTH_SOCK llmgw server init \\\n    --patroni <user>@<a> --patroni <user>@<b> --no-mirror --no-node --no-model"
+	case cJudge:
+		item = "\n  Serves the model the guardrail policies ask as their judge — Llama\n  Guard 3 1B on CPUs, under the name the policies give it — so what they\n  judge is not refused (guard_unavailable). A model of that name that is\n  not a judge is replaced. Overview says when a judge is missing.\n\n  " +
+			sDim.Render("On the command line:") + "\n  llmgw ctl guard judge status\n  llmgw ctl guard judge enable --replace"
 	case cUpgrade:
 		item = "\n  Moves the cluster to a release: this machine first, then each node\n  in turn. Engines keep serving; a node that does not come back stops it.\n\n  " +
 			sDim.Render("On the command line:") + "\n  sudo llmgw ctl upgrade VERSION"

internal/cli/tui_kb.go

@@ -0,0 +1,242 @@
+package cli
+
+import (
+	"context"
+	"encoding/json"
+	"fmt"
+	"net/http"
+	"net/url"
+	"strconv"
+	"strings"
+	"time"
+
+	tea "github.com/charmbracelet/bubbletea"
+
+	"gitea.mynodes.xyz/ilker/llmgw/internal/control/adminapi"
+)
+
+// Knowledge bases, as the console shows them: each collection with its
+// owner, model, documents and grants; Enter for what can be done to one, A
+// to make one. People fill them from their own machines with llmgw-kb.
+
+type kbState struct {
+	view adminapi.KBView
+	err  error
+	// loaded is set once an answer came, so "none yet" is not shown while
+	// asking.
+	loaded bool
+}
+
+type consoleKBMsg struct {
+	view adminapi.KBView
+	err  error
+}
+
+func (m consoleModel) fetchKB() tea.Cmd {
+	c := m.client
+	return func() tea.Msg {
+		ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
+		defer cancel()
+		var msg consoleKBMsg
+		msg.err = c.get(ctx, "/api/v1/kb", &msg.view)
+		return msg
+	}
+}
+
+func (m consoleModel) kbPanel(pw int) (string, string, string) {
+	keys := "→ or Enter open · ↑↓ choose · Q leave"
+	if m.inPanel {
+		keys = "↑↓ collection · Enter actions · A new collection · ← back · Q leave"
+	}
+	k := m.kb
+	var b strings.Builder
+	b.WriteString("\n  " + sDim.Render("Documents requests are answered from; people add them with llmgw-kb.") + "\n\n")
+	switch {
+	case k.err != nil:
+		b.WriteString("  " + sBad.Render(k.err.Error()))
+		return "Knowledge", b.String(), keys
+	case !k.loaded:
+		b.WriteString("  " + sDim.Render("asking the control plane…"))
+		return "Knowledge", b.String(), keys
+	case !k.view.Available:
+		b.WriteString("  " + sBad.Render("Off: ") + k.view.Reason)
+		return "Knowledge", b.String(), keys
+	case len(k.view.Collections) == 0:
+		b.WriteString("  " + sDim.Render("No collections yet. A to make one."))
+		return "Knowledge", b.String(), keys
+	}
+	var rows [][]string
+	for _, c := range k.view.Collections {
+		var grants []string
+		for _, g := range c.Grants {
+			grants = append(grants, or(g.Name, g.Target)+" "+g.Perm)
+		}
+		rows = append(rows, []string{or(c.OrgName, c.Org) + "/" + c.Name, or(c.OwnerName, c.Owner),
+			fmt.Sprintf("%s (%d)", c.EmbedModel, c.Dims), fmt.Sprint(c.Documents), fmt.Sprint(c.Chunks), dash(strings.Join(grants, ", "))})
+	}
+	b.WriteString(table([]string{"COLLECTION", "OWNER", "MODEL", "DOCUMENTS", "CHUNKS", "GRANTED TO"}, rows, m.row, pw, m.inPanel))
+	if len(k.view.Defaults) > 0 {
+		b.WriteString("\n\n  " + sTitle.Render("BY DEFAULT") + "  " + sDim.Render("when a request names no collection") + "\n")
+		var drows [][]string
+		for _, d := range k.view.Defaults {
+			drows = append(drows, []string{or(d.Name, d.Target), strings.Join(d.Collections, ", ")})
+		}
+		b.WriteString(table([]string{"FOR", "ANSWERED FROM"}, drows, -1, pw, false))
+	}
+	b.WriteString("\n\n  " + sDim.Render("On a person's machine: llmgw-kb sync COLLECTION FOLDER"))
+	return "Knowledge", b.String(), keys
+}
+
+// kbRef is how the admin API is told which collection: by its ID.
+func kbRef(c adminapi.KBCollectionView) string { return url.PathEscape(c.ID) }
+
+// kbCreateForm makes a collection.
+func (m *consoleModel) kbCreateForm() {
+	model := ""
+	for _, md := range m.data.models {
+		n := strings.ToLower(md.Name)
+		if strings.Contains(n, "embed") || strings.Contains(n, "bge") || strings.Contains(n, "e5") {
+			model = md.Name
+			break
+		}
+	}
+	m.openForm(&formDialog{title: "New knowledge-base collection",
+		note: "Owned by part of the tree: it and everything below it may read and add to it.\n" +
+			"The model must embed (capabilities.embeddings) and be able to serve now.",
+		fields: []formField{
+			choiceF("owner", []string{"project", "team", "org", "user"}, "team"),
+			textF("which", "", "ORG/TEAM/PROJECT, ORG/TEAM, ORG, or the ID"),
+			textF("org", "", "for a person's collection only"),
+			textF("name", "", "how people name it to llmgw-kb"),
+			textF("model", model, "the catalog model that embeds it"),
+		},
+		submit: func(m *consoleModel, v []string) (tea.Cmd, error) {
+			for i, what := range map[int]string{1: "which " + v[0], 3: "a name", 4: "a model"} {
+				if strings.TrimSpace(v[i]) == "" {
+					return nil, fmt.Errorf("%s?", what)
+				}
+			}
+			req := adminapi.KBCreateRequest{Owner: v[0] + ":" + v[1], Org: v[2], Name: v[3], Model: v[4]}
+			return m.call("New collection "+v[3], http.MethodPost, "/api/v1/kb/collections", req, func(raw json.RawMessage) string {
+				var c adminapi.KBCollectionView
+				_ = json.Unmarshal(raw, &c)
+				return fmt.Sprintf("Collection %s, embedded by %s in %d dimensions.\n\nPeople with access fill it with:\n  llmgw-kb sync %s FOLDER",
+					c.Name, c.EmbedModel, c.Dims, c.Name)
+			}), nil
+		}})
+}
+
+// kbMenu is what can be done to one collection.
+func (m consoleModel) kbMenu(c adminapi.KBCollectionView) *menuDialog {
+	name := or(c.OrgName, c.Org) + "/" + c.Name
+	items := []menuItem{
+		{"Documents", func(m *consoleModel) tea.Cmd { return m.kbDocuments(c) }},
+		{"Grant…", func(m *consoleModel) tea.Cmd {
+			m.openForm(&formDialog{title: "Grant " + name,
+				note: "It and everything below it may then read, or with write add to, this collection.\nNever outside " + or(c.OrgName, c.Org) + ".",
+				fields: []formField{
+					choiceF("to", []string{"project", "team", "org", "user", "key"}, "project"),
+					textF("which", "", "ORG/TEAM/PROJECT, ORG/TEAM, ORG, or the ID"),
+					choiceF("may", []string{"read", "write"}, "read"),
+				},
+				submit: func(m *consoleModel, v []string) (tea.Cmd, error) {
+					if strings.TrimSpace(v[1]) == "" {
+						return nil, fmt.Errorf("which %s?", v[0])
+					}
+					req := adminapi.KBGrantRequest{Collection: c.ID, Target: v[0] + ":" + strings.TrimSpace(v[1]), Perm: v[2]}
+					return m.call("Grant "+c.Name, http.MethodPost, "/api/v1/kb/grants", req, nil), nil
+				}})
+			return nil
+		}},
+	}
+	for _, g := range c.Grants {
+		items = append(items, menuItem{"Revoke " + or(g.Name, g.Target) + " (" + g.Perm + ")…", func(m *consoleModel) tea.Cmd {
+			m.confirmD = &confirmDialog{title: "Revoke a grant on " + name,
+				text: or(g.Name, g.Target) + " may then no longer " + g.Perm + " it, unless it owns it.",
+				yes: func(m *consoleModel) tea.Cmd {
+					return m.call("Revoke", http.MethodDelete, "/api/v1/kb/grants",
+						adminapi.KBGrantRequest{Collection: c.ID, Target: g.Target}, nil)
+				}}
+			return nil
+		}})
+	}
+	items = append(items,
+		menuItem{"Answer from it by default…", func(m *consoleModel) tea.Cmd {
+			m.openForm(&formDialog{title: "Answer from " + name + " by default",
+				note: "Chat requests from there that name no collection are answered from this one.\n" +
+					"Replaces that part's default; llmgw ctl kb default sets several.",
+				fields: []formField{
+					choiceF("for", []string{"project", "team", "org", "key"}, "team"),
+					textF("which", "", "ORG/TEAM/PROJECT, ORG/TEAM, ORG, or the ID"),
+				},
+				submit: func(m *consoleModel, v []string) (tea.Cmd, error) {
+					if strings.TrimSpace(v[1]) == "" {
+						return nil, fmt.Errorf("which %s?", v[0])
+					}
+					req := adminapi.KBDefaultRequest{Target: v[0] + ":" + strings.TrimSpace(v[1]), Collections: []string{c.ID}}
+					return m.call("Default "+c.Name, http.MethodPost, "/api/v1/kb/defaults", req, nil), nil
+				}})
+			return nil
+		}},
+		menuItem{"Try a search…", func(m *consoleModel) tea.Cmd {
+			m.openForm(&formDialog{title: "Search " + name,
+				note:   "As an administrator: no tenancy applies. Shows what a question would find.",
+				fields: []formField{textF("question", "", ""), textF("results", "5", "how many chunks")},
+				submit: func(m *consoleModel, v []string) (tea.Cmd, error) {
+					if v[0] == "" {
+						return nil, fmt.Errorf("what question?")
+					}
+					k, _ := strconv.Atoi(v[1])
+					req := adminapi.KBSearchRequest{Collection: c.ID, Query: v[0], K: k}
+					return m.call("Search "+c.Name, http.MethodPost, "/api/v1/kb/search", req, func(raw json.RawMessage) string {
+						var hits []adminapi.KBHitView
+						_ = json.Unmarshal(raw, &hits)
+						if len(hits) == 0 {
+							return "Nothing found."
+						}
+						var b strings.Builder
+						for i, h := range hits {
+							text := h.Text
+							if r := []rune(text); len(r) > 400 {
+								text = string(r[:400]) + "…"
+							}
+							fmt.Fprintf(&b, "%d. %s #%d\n%s\n\n", i+1, h.Document, h.Ord, indent(strings.TrimSpace(text), "   "))
+						}
+						return strings.TrimRight(b.String(), "\n")
+					}), nil
+				}})
+			return nil
+		}},
+		menuItem{"Delete…", func(m *consoleModel) tea.Cmd {
+			m.confirmD = &confirmDialog{title: "Delete " + name,
+				text: fmt.Sprintf("Its %d documents and %d chunks, and every grant, go with it. People's own files are not touched.",
+					c.Documents, c.Chunks),
+				yes: func(m *consoleModel) tea.Cmd {
+					return m.call("Delete "+c.Name, http.MethodDelete, "/api/v1/kb/collections/"+kbRef(c), nil, nil)
+				}}
+			return nil
+		}},
+	)
+	return &menuDialog{title: name, items: items}
+}
+
+// kbDocuments shows a collection's documents in a dialog.
+func (m *consoleModel) kbDocuments(c adminapi.KBCollectionView) tea.Cmd {
+	return m.call("Documents of "+c.Name, http.MethodGet, "/api/v1/kb/collections/"+kbRef(c)+"/documents", nil,
+		func(raw json.RawMessage) string {
+			var docs []adminapi.KBDocumentView
+			_ = json.Unmarshal(raw, &docs)
+			if len(docs) == 0 {
+				return "None yet. On a person's machine: llmgw-kb sync " + c.Name + " FOLDER"
+			}
+			var b strings.Builder
+			for i, d := range docs {
+				if i == 40 {
+					fmt.Fprintf(&b, "… and %d more (llmgw ctl kb docs %s)", len(docs)-40, c.ID)
+					break
+				}
+				fmt.Fprintf(&b, "%-48s %5d chunks  %s  %s\n", d.Name, d.Chunks, d.AddedAt.Local().Format("2006-01-02 15:04"), d.AddedBy)
+			}
+			return strings.TrimRight(b.String(), "\n")
+		})
+}

internal/cli/tui_kb_test.go

@@ -0,0 +1,62 @@
+package cli
+
+import (
+	"encoding/json"
+	"strings"
+	"testing"
+
+	tea "github.com/charmbracelet/bubbletea"
+
+	"gitea.mynodes.xyz/ilker/llmgw/internal/control/adminapi"
+)
+
+func TestACollectionIsMadeAndGrantedFromTheConsole(t *testing.T) {
+	m, s := consoleWith(t, cKB)
+	m.data.models = append(m.data.models, adminapi.ModelView{Name: "bge-m3"})
+	m.kb = kbState{loaded: true, view: adminapi.KBView{Available: true, Collections: []adminapi.KBCollectionView{{
+		ID: "kb_aaaaaaaaaaaaaaaa", Name: "policies", Org: "o1", OrgName: "Acme", Owner: "team:t1", OwnerName: "Acme / hr",
+		EmbedModel: "bge-m3", Dims: 1024, Documents: 3, Chunks: 40,
+		Grants: []adminapi.KBGrantView{{Target: "project:p9", Name: "Acme/sales/crm", Perm: "read"}},
+	}}}}
+	var tm tea.Model = m
+	view := tm.View()
+	for _, want := range []string{"Knowledge", "Acme/policies", "Acme / hr", "bge-m3 (1024)", "Acme/s", "llmgw-kb sync COLLECTION FOLDER"} {
+		if !strings.Contains(view, want) {
+			t.Errorf("missing %q in:\n%s", want, view)
+		}
+	}
+
+	// A: the form, prefilled with the embedding model.
+	tm = press(tm, "a")
+	f := tm.(consoleModel).formD
+	if f == nil || f.fields[4].value() != "bge-m3" {
+		t.Fatalf("no form, or the model not offered: %+v", f)
+	}
+	f.fields[1].text.SetValue("Acme/hr")
+	f.fields[3].text.SetValue("handbook")
+	tm = pressRun(tm, "tab", "tab", "tab", "tab", "tab", "enter")
+	var req adminapi.KBCreateRequest
+	got := s.last()
+	if !strings.HasPrefix(got, "POST /api/v1/kb/collections ") ||
+		json.Unmarshal([]byte(strings.SplitN(got, " ", 3)[2]), &req) != nil ||
+		req.Owner != "team:Acme/hr" || req.Name != "handbook" || req.Model != "bge-m3" {
+		t.Fatalf("sent %q", got)
+	}
+
+	// Enter on a row: grant to a project for writing.
+	m.formD = nil
+	tm = m
+	tm = press(tm, "enter")
+	if md := tm.(consoleModel).menuD; md == nil || !strings.Contains(md.items[2].label, "Revoke Acme/sales/crm") {
+		t.Fatalf("menu %+v", md)
+	}
+	tm = press(tm, "down", "enter")
+	f = tm.(consoleModel).formD
+	f.fields[1].text.SetValue("Acme/sales/crm")
+	f.fields[2].opt = 1 // write
+	tm = pressRun(tm, "tab", "tab", "tab", "enter")
+	if got := s.last(); got != `POST /api/v1/kb/grants {"collection":"kb_aaaaaaaaaaaaaaaa","target":"project:Acme/sales/crm","perm":"write"}` {
+		t.Fatalf("sent %q", got)
+	}
+	_ = tm
+}

internal/cli/users.go

@@ -3,6 +3,7 @@ package cli
 import (
 	"errors"
 	"fmt"
+	"net/url"
 	"strings"
 	"text/tabwriter"
 	"time"
@@ -34,7 +35,7 @@ func newUsersCmd() *cobra.Command {
 			"Stop an account working, keeping it and its history"),
 		newUserStateCmd(open, "reinstate", store.UserActive,
 			"Let a suspended account work again"),
-		newUserCreateCmd(open), newUserResetPasswordCmd(open),
+		newUserCreateCmd(open), newUserResetPasswordCmd(open), newUserResetLinkCmd(open, &sf),
 		newUserJoinCmd(open), newUserLeaveCmd(open), newUserDeleteCmd(open), newUserForgetCmd(open),
 		newUserAdminCmd(open, "promote", true, "Let an account see the cluster dashboard at /admin"),
 		newUserAdminCmd(open, "demote", false, "Take away access to the cluster dashboard"),
@@ -254,9 +255,10 @@ func newUserResetPasswordCmd(open userOpener) *cobra.Command {
 	cmd := &cobra.Command{
 		Use:   "reset-password <user>",
 		Short: "Set a new password for someone who lost theirs",
-		Long: "There is no self-service reset, because this gateway sends no email and\n" +
-			"a reset link with nowhere to send it is not a reset. Set one here, tell\n" +
-			"them out of band, and have them change it from their profile.\n\n" +
+		Long: "People reset a forgotten password themselves at /forgot when the cluster\n" +
+			"has mail settings (mail.smtp); without them, 'llmgw users reset-link' makes\n" +
+			"a link to hand them. Or set one here, tell them out of band, and have them\n" +
+			"change it from their profile.\n\n" +
 			"Every session they have open is closed: if they lost the password, the\n" +
 			"reason might be that somebody else has it.",
 		Args: cobra.ExactArgs(1),
@@ -532,3 +534,50 @@ func newUserForgetCmd(open userOpener) *cobra.Command {
 	cmd.Flags().BoolVar(&yes, "yes", false, "confirm")
 	return cmd
 }
+
+func newUserResetLinkCmd(open userOpener, sf *storeFlags) *cobra.Command {
+	var base string
+	cmd := &cobra.Command{
+		Use:   "reset-link <user>",
+		Short: "Make a one-time link for someone to set a forgotten password",
+		Long: "For a gateway that sends no mail, or a person whose mail does not arrive:\n" +
+			"the link is printed, to hand them however suits. It works once, for an\n" +
+			"hour, and a newer one replaces it. Setting the password closes every\n" +
+			"session they have open. The address it points at is mail.base_url from\n" +
+			"the cluster's settings, or --base-url.",
+		Args: cobra.ExactArgs(1),
+		RunE: func(cmd *cobra.Command, args []string) error {
+			if base == "" {
+				if cfg, err := loadEffectiveConfig(sf.config); err == nil {
+					base = cfg.Mail.BaseURL
+				}
+			}
+			if base == "" {
+				return errors.New("give --base-url, the address people reach this gateway at, or set mail.base_url " +
+					"(llmgw ctl config set mail.base_url https://…)")
+			}
+			s, err := open(cmd)
+			if err != nil {
+				return err
+			}
+			defer s.Close()
+			u, err := resolveUser(cmd.Context(), s, args[0])
+			if err != nil {
+				return err
+			}
+			if u.PasswordHash == "" {
+				return fmt.Errorf("%s signs in through single sign-on: there is no password to set", u.Email)
+			}
+			token, err := s.CreatePasswordReset(cmd.Context(), u.ID, account.ResetLinkTTL)
+			if err != nil {
+				return err
+			}
+			fmt.Fprintf(cmd.OutOrStdout(), "%s/reset?token=%s\n", strings.TrimRight(base, "/"), url.QueryEscape(token))
+			cmd.PrintErrf("For %s: works once, for %s. Hand it to them alone; it opens their account.\n",
+				u.Email, account.ResetLinkTTL)
+			return nil
+		},
+	}
+	cmd.Flags().StringVar(&base, "base-url", "", "the address people reach this gateway at (default: mail.base_url)")
+	return cmd
+}

internal/comfy/comfy.go

@@ -0,0 +1,529 @@
+// Package comfy has images made by ComfyUI: a workflow filled in with what
+// was asked for, queued, waited on, and its pictures fetched.
+//
+// ComfyUI speaks no OpenAI. A workflow is a graph of nodes that is queued with
+// POST /prompt, is done when GET /history/{id} lists it, and leaves its
+// pictures to be read one by one from GET /view. What the graph is depends on
+// the model — FLUX schnell is not FLUX dev, and neither is SDXL — so the graph
+// is a template the catalog names, with placeholders where a request's
+// prompt, size and seed go.
+package comfy
+
+import (
+	"bytes"
+	"context"
+	"crypto/rand"
+	"embed"
+	"encoding/binary"
+	"encoding/hex"
+	"encoding/json"
+	"errors"
+	"fmt"
+	"io"
+	"math/big"
+	"net/http"
+	"net/url"
+	"slices"
+	"strconv"
+	"strings"
+	"time"
+)
+
+//go:embed workflows/*.json
+var builtin embed.FS
+
+// DefaultWorkflow is the template used when the catalog names none.
+const DefaultWorkflow = "flux-schnell"
+
+// Builtins are the workflows that come with llmgw, by name.
+func Builtins() []string {
+	entries, _ := builtin.ReadDir("workflows")
+	out := make([]string, 0, len(entries))
+	for _, e := range entries {
+		out = append(out, strings.TrimSuffix(e.Name(), ".json"))
+	}
+	return out
+}
+
+// Template is a workflow in ComfyUI's API format, with placeholders.
+//
+// A placeholder is a string value "{{name}}", or "{{name|default}}" for one a
+// request may leave out. One that is the whole value becomes the value
+// itself, a number where the node wants a number; one inside a longer string
+// is written into it. The names are those of Params.
+type Template struct {
+	graph map[string]any
+}
+
+// Load reads a template: a built-in's name, or the workflow's JSON itself.
+func Load(nameOrJSON string) (*Template, error) {
+	s := strings.TrimSpace(nameOrJSON)
+	if s == "" {
+		s = DefaultWorkflow
+	}
+	raw := []byte(s)
+	if !strings.HasPrefix(s, "{") {
+		b, err := builtin.ReadFile("workflows/" + s + ".json")
+		if err != nil {
+			return nil, fmt.Errorf("no built-in workflow %q (there are %s)", s, strings.Join(Builtins(), ", "))
+		}
+		raw = b
+	}
+	var g map[string]any
+	if err := json.Unmarshal(raw, &g); err != nil {
+		return nil, fmt.Errorf("the workflow is not JSON: %w", err)
+	}
+	if len(g) == 0 {
+		return nil, errors.New("the workflow has no nodes")
+	}
+	for id, n := range g {
+		node, ok := n.(map[string]any)
+		if !ok || node["class_type"] == nil {
+			// The UI's own save format has a "nodes" list instead, and is
+			// not something /prompt accepts.
+			return nil, fmt.Errorf("node %q has no class_type: the workflow must be saved in API format", id)
+		}
+	}
+	return &Template{graph: g}, nil
+}
+
+// Params are what a request decides. A zero number is one not given, which
+// leaves it to the template's default.
+type Params struct {
+	Checkpoint     string
+	Prompt         string
+	NegativePrompt string
+	Width, Height  int
+	N              int
+	Seed           int64
+	Steps          int
+	Guidance       float64
+}
+
+func (p Params) value(name string) (any, bool) {
+	switch name {
+	case "checkpoint":
+		return p.Checkpoint, p.Checkpoint != ""
+	case "prompt":
+		return p.Prompt, true
+	case "negative_prompt":
+		return p.NegativePrompt, true
+	case "width":
+		return p.Width, p.Width > 0
+	case "height":
+		return p.Height, p.Height > 0
+	case "n":
+		return max(p.N, 1), true
+	case "seed":
+		return p.Seed, true
+	case "steps":
+		return p.Steps, p.Steps > 0
+	case "guidance":
+		return p.Guidance, p.Guidance > 0
+	}
+	return nil, false
+}
+
+// Fill is the workflow for one request.
+func (t *Template) Fill(p Params) (map[string]any, error) {
+	var missing []string
+	var fill func(v any) any
+	fill = func(v any) any {
+		switch v := v.(type) {
+		case map[string]any:
+			out := make(map[string]any, len(v))
+			for k, x := range v {
+				out[k] = fill(x)
+			}
+			return out
+		case []any:
+			out := make([]any, len(v))
+			for i, x := range v {
+				out[i] = fill(x)
+			}
+			return out
+		case string:
+			return fillString(v, p, &missing)
+		}
+		return v
+	}
+	g := fill(t.graph).(map[string]any)
+	if len(missing) > 0 {
+		slices.Sort(missing)
+		return nil, fmt.Errorf("the workflow wants %s, which the request did not give and the workflow has no default for",
+			strings.Join(slices.Compact(missing), ", "))
+	}
+	return g, nil
+}
+
+func fillString(s string, p Params, missing *[]string) any {
+	if !strings.Contains(s, "{{") {
+		return s
+	}
+	resolve := func(ph string) (any, bool) {
+		name, def, hasDef := strings.Cut(ph, "|")
+		name = strings.TrimSpace(name)
+		if v, ok := p.value(name); ok {
+			return v, true
+		}
+		if hasDef {
+			def = strings.TrimSpace(def)
+			if f, err := strconv.ParseFloat(def, 64); err == nil {
+				return f, true
+			}
+			return def, true
+		}
+		*missing = append(*missing, name)
+		return nil, false
+	}
+	// The whole value: what the node gets is the value, not its text.
+	if strings.HasPrefix(s, "{{") && strings.HasSuffix(s, "}}") && strings.Count(s, "{{") == 1 {
+		v, _ := resolve(s[2 : len(s)-2])
+		return v
+	}
+	var b strings.Builder
+	for {
+		i := strings.Index(s, "{{")
+		j := strings.Index(s, "}}")
+		if i < 0 || j < i {
+			b.WriteString(s)
+			return b.String()
+		}
+		b.WriteString(s[:i])
+		if v, ok := resolve(s[i+2 : j]); ok {
+			fmt.Fprint(&b, v)
+		}
+		s = s[j+2:]
+	}
+}
+
+// RandomSeed is a seed for a request that named none. ComfyUI's are 64-bit,
+// but JSON numbers are read as doubles on the way: kept to 2^53.
+func RandomSeed() int64 {
+	n, err := rand.Int(rand.Reader, big.NewInt(1<<53))
+	if err != nil {
+		return time.Now().UnixNano() & (1<<53 - 1)
+	}
+	return n.Int64()
+}
+
+// Image is one picture made.
+type Image struct {
+	Data        []byte
+	ContentType string
+}
+
+// Error is a failure, said as what it means for the request.
+type Error struct {
+	Msg string
+	// Retryable is a failure of this server, which another one might not
+	// have: it could not be reached, or it answered 5xx.
+	Retryable bool
+}
+
+func (e *Error) Error() string { return e.Msg }
+
+// Client queues workflows and waits for them.
+type Client struct {
+	HTTP *http.Client
+	// Poll is how often a queued workflow is asked after.
+	Poll time.Duration
+}
+
+// Generate runs a workflow on the ComfyUI at base and returns its pictures,
+// in the order of the nodes that made them. A context that ends takes the
+// workflow out of the queue, or interrupts it if it is running, so a client
+// that went away does not leave the card working for nobody.
+func (c *Client) Generate(ctx context.Context, base, apiKey string, workflow map[string]any) ([]Image, error) {
+	base = strings.TrimRight(base, "/")
+	client := make([]byte, 8)
+	_, _ = rand.Read(client)
+	body, err := json.Marshal(map[string]any{"prompt": workflow, "client_id": "llmgw-" + hex.EncodeToString(client)})
+	if err != nil {
+		return nil, err
+	}
+	var queued struct {
+		PromptID   string          `json:"prompt_id"`
+		Error      json.RawMessage `json:"error"`
+		NodeErrors json.RawMessage `json:"node_errors"`
+	}
+	status, err := c.do(ctx, http.MethodPost, base+"/prompt", apiKey, body, &queued)
+	if err != nil {
+		return nil, err
+	}
+	if status != http.StatusOK || queued.PromptID == "" {
+		// A workflow ComfyUI will not run: a checkpoint it does not have, a
+		// node it does not know. The catalog's to fix, not the client's, and
+		// no other replica set up the same way would take it either.
+		return nil, &Error{Msg: "ComfyUI refused the workflow: " + refusal(queued.Error, queued.NodeErrors)}
+	}
+	id := queued.PromptID
+
+	poll := c.Poll
+	if poll <= 0 {
+		poll = 250 * time.Millisecond
+	}
+	t := time.NewTicker(poll)
+	defer t.Stop()
+	for {
+		select {
+		case <-ctx.Done():
+			c.cancel(context.WithoutCancel(ctx), base, apiKey, id)
+			return nil, ctx.Err()
+		case <-t.C:
+		}
+		var hist map[string]historyEntry
+		if _, err := c.do(ctx, http.MethodGet, base+"/history/"+url.PathEscape(id), apiKey, nil, &hist); err != nil {
+			if ctx.Err() != nil {
+				continue // said by the select
+			}
+			return nil, err
+		}
+		h, ok := hist[id]
+		if !ok {
+			continue
+		}
+		if h.Status.StatusStr == "error" {
+			return nil, &Error{Msg: "the workflow failed in ComfyUI: " + h.failure()}
+		}
+		if !h.Status.Completed {
+			continue
+		}
+		return c.fetch(ctx, base, apiKey, h)
+	}
+}
+
+type historyEntry struct {
+	Status struct {
+		StatusStr string              `json:"status_str"`
+		Completed bool                `json:"completed"`
+		Messages  [][]json.RawMessage `json:"messages"`
+	} `json:"status"`
+	Outputs map[string]struct {
+		Images []struct {
+			Filename  string `json:"filename"`
+			Subfolder string `json:"subfolder"`
+			Type      string `json:"type"`
+		} `json:"images"`
+	} `json:"outputs"`
+}
+
+// failure is what ComfyUI said went wrong, from its execution_error.
+func (h historyEntry) failure() string {
+	for _, m := range h.Status.Messages {
+		if len(m) < 2 || string(m[0]) != `"execution_error"` {
+			continue
+		}
+		var e struct {
+			NodeType string `json:"node_type"`
+			Message  string `json:"exception_message"`
+		}
+		if json.Unmarshal(m[1], &e) == nil && e.Message != "" {
+			return strings.TrimSpace(e.NodeType + ": " + strings.TrimSpace(e.Message))
+		}
+	}
+	return "no reason given"
+}
+
+func (c *Client) fetch(ctx context.Context, base, apiKey string, h historyEntry) ([]Image, error) {
+	nodes := make([]string, 0, len(h.Outputs))
+	for id := range h.Outputs {
+		nodes = append(nodes, id)
+	}
+	// Numerically where they are numbers, which is how ComfyUI names them.
+	slices.SortFunc(nodes, func(a, b string) int {
+		x, ea := strconv.Atoi(a)
+		y, eb := strconv.Atoi(b)
+		if ea == nil && eb == nil {
+			return x - y
+		}
+		return strings.Compare(a, b)
+	})
+	var out []Image
+	for _, n := range nodes {
+		for _, im := range h.Outputs[n].Images {
+			q := url.Values{"filename": {im.Filename}, "subfolder": {im.Subfolder}, "type": {im.Type}}
+			img, err := c.get(ctx, base+"/view?"+q.Encode(), apiKey)
+			if err != nil {
+				return nil, err
+			}
+			out = append(out, img)
+		}
+	}
+	if len(out) == 0 {
+		return nil, &Error{Msg: "the workflow finished without a picture: it has no SaveImage or PreviewImage node"}
+	}
+	return out, nil
+}
+
+// cancel takes a workflow out of the queue, or stops it if it is the one
+// running. Best effort: the server may be gone, which also stops it.
+func (c *Client) cancel(ctx context.Context, base, apiKey, id string) {
+	ctx, done := context.WithTimeout(ctx, 5*time.Second)
+	defer done()
+	body, _ := json.Marshal(map[string]any{"delete": []string{id}})
+	_, _ = c.do(ctx, http.MethodPost, base+"/queue", apiKey, body, nil)
+	var q struct {
+		Running [][]json.RawMessage `json:"queue_running"`
+	}
+	if _, err := c.do(ctx, http.MethodGet, base+"/queue", apiKey, nil, &q); err != nil {
+		return
+	}
+	for _, r := range q.Running {
+		if len(r) > 1 && string(r[1]) == strconv.Quote(id) {
+			_, _ = c.do(ctx, http.MethodPost, base+"/interrupt", apiKey, []byte("{}"), nil)
+			return
+		}
+	}
+}
+
+func (c *Client) client() *http.Client {
+	if c.HTTP != nil {
+		return c.HTTP
+	}
+	return http.DefaultClient
+}
+
+func (c *Client) request(ctx context.Context, method, u, apiKey string, body []byte) (*http.Response, error) {
+	req, err := http.NewRequestWithContext(ctx, method, u, bytes.NewReader(body))
+	if err != nil {
+		return nil, err
+	}
+	if body != nil {
+		req.Header.Set("Content-Type", "application/json")
+	}
+	if apiKey != "" {
+		req.Header.Set("Authorization", "Bearer "+apiKey)
+	}
+	resp, err := c.client().Do(req)
+	if err != nil {
+		if ctx.Err() != nil {
+			return nil, ctx.Err()
+		}
+		return nil, &Error{Msg: "ComfyUI could not be reached: " + err.Error(), Retryable: true}
+	}
+	return resp, nil
+}
+
+// do sends a request and reads a JSON answer into out. A 5xx is an error;
+// any other status is returned with whatever body could be read.
+func (c *Client) do(ctx context.Context, method, u, apiKey string, body []byte, out any) (int, error) {
+	resp, err := c.request(ctx, method, u, apiKey, body)
+	if err != nil {
+		return 0, err
+	}
+	defer resp.Body.Close()
+	b, err := io.ReadAll(io.LimitReader(resp.Body, 16<<20))
+	if err != nil {
+		return 0, &Error{Msg: "reading ComfyUI's answer: " + err.Error(), Retryable: true}
+	}
+	if resp.StatusCode >= 500 {
+		return resp.StatusCode, &Error{Msg: fmt.Sprintf("ComfyUI answered %d: %s", resp.StatusCode, snippet(b)), Retryable: true}
+	}
+	if out != nil && len(b) > 0 {
+		if err := json.Unmarshal(b, out); err != nil && resp.StatusCode == http.StatusOK {
+			return resp.StatusCode, &Error{Msg: "ComfyUI's answer is not JSON: " + snippet(b)}
+		}
+	}
+	return resp.StatusCode, nil
+}
+
+func (c *Client) get(ctx context.Context, u, apiKey string) (Image, error) {
+	resp, err := c.request(ctx, http.MethodGet, u, apiKey, nil)
+	if err != nil {
+		return Image{}, err
+	}
+	defer resp.Body.Close()
+	b, err := io.ReadAll(io.LimitReader(resp.Body, 64<<20))
+	if err != nil {
+		return Image{}, &Error{Msg: "reading a picture from ComfyUI: " + err.Error(), Retryable: true}
+	}
+	if resp.StatusCode != http.StatusOK {
+		return Image{}, &Error{Msg: fmt.Sprintf("ComfyUI answered %d for a picture it made", resp.StatusCode)}
+	}
+	ct := resp.Header.Get("Content-Type")
+	if ct == "" || ct == "application/octet-stream" {
+		ct = http.DetectContentType(b)
+	}
+	if ct == "image/png" {
+		b = withoutPNGText(b)
+	}
+	return Image{Data: b, ContentType: ct}, nil
+}
+
+// withoutPNGText is a PNG without its text chunks. ComfyUI writes the whole
+// workflow into every picture it saves — the prompt, and the checkpoint and
+// nodes behind it — and what a picture says about how it was made is the
+// operator's to publish, not something to hand every client. A file that is
+// not a well-formed PNG is returned as it came.
+func withoutPNGText(b []byte) []byte {
+	const sig = "\x89PNG\r\n\x1a\n"
+	if len(b) < len(sig) || string(b[:len(sig)]) != sig {
+		return b
+	}
+	out := make([]byte, 0, len(b))
+	out = append(out, b[:len(sig)]...)
+	for i := len(sig); i < len(b); {
+		if i+12 > len(b) {
+			return b
+		}
+		n := int(binary.BigEndian.Uint32(b[i:]))
+		end := i + 12 + n
+		if n < 0 || end > len(b) {
+			return b
+		}
+		switch string(b[i+4 : i+8]) {
+		case "tEXt", "iTXt", "zTXt":
+		default:
+			out = append(out, b[i:end]...)
+		}
+		i = end
+	}
+	return out
+}
+
+// refusal is what /prompt said when it would not queue a workflow.
+func refusal(e, nodes json.RawMessage) string {
+	var top struct {
+		Message string `json:"message"`
+		Details string `json:"details"`
+	}
+	msg := snippet(e)
+	if json.Unmarshal(e, &top) == nil && top.Message != "" {
+		msg = top.Message
+		if top.Details != "" {
+			msg += ": " + top.Details
+		}
+	}
+	var perNode map[string]struct {
+		ClassType string `json:"class_type"`
+		Errors    []struct {
+			Message string `json:"message"`
+			Details string `json:"details"`
+		} `json:"errors"`
+	}
+	if json.Unmarshal(nodes, &perNode) == nil {
+		var parts []string
+		for _, n := range perNode {
+			for _, er := range n.Errors {
+				parts = append(parts, strings.TrimSpace(n.ClassType+": "+er.Message+" "+er.Details))
+			}
+		}
+		slices.Sort(parts)
+		if len(parts) > 0 {
+			msg += " (" + strings.Join(parts, "; ") + ")"
+		}
+	}
+	if msg == "" {
+		msg = "no reason given"
+	}
+	return msg
+}
+
+func snippet(b []byte) string {
+	s := strings.TrimSpace(string(b))
+	if len(s) > 300 {
+		s = s[:300] + "…"
+	}
+	return s
+}

internal/comfy/comfy_test.go

@@ -0,0 +1,227 @@
+package comfy
+
+import (
+	"context"
+	"encoding/binary"
+	"encoding/json"
+	"fmt"
+	"net/http"
+	"net/http/httptest"
+	"strings"
+	"sync"
+	"testing"
+	"time"
+)
+
+func TestTheBuiltInWorkflowsFillIn(t *testing.T) {
+	for _, name := range Builtins() {
+		tpl, err := Load(name)
+		if err != nil {
+			t.Fatalf("%s: %v", name, err)
+		}
+		g, err := tpl.Fill(Params{Checkpoint: "flux.safetensors", Prompt: `a "red" fox`, Width: 768, Height: 512, N: 2, Seed: 42})
+		if err != nil {
+			t.Fatalf("%s: %v", name, err)
+		}
+		b, _ := json.Marshal(g)
+		s := string(b)
+		for _, want := range []string{`"ckpt_name":"flux.safetensors"`, `"text":"a \"red\" fox"`, `"width":768`,
+			`"height":512`, `"batch_size":2`, `"seed":42`} {
+			if !strings.Contains(s, want) {
+				t.Errorf("%s: %s not in %s", name, want, s)
+			}
+		}
+		if strings.Contains(s, "{{") {
+			t.Errorf("%s: a placeholder was left: %s", name, s)
+		}
+	}
+}
+
+func TestAStepCountGivenReplacesTheDefault(t *testing.T) {
+	tpl, _ := Load("flux-schnell")
+	g, _ := tpl.Fill(Params{Checkpoint: "c", Prompt: "p", Width: 512, Height: 512})
+	if steps := g["5"].(map[string]any)["inputs"].(map[string]any)["steps"]; steps != 4.0 {
+		t.Errorf("default steps = %v, want 4", steps)
+	}
+	g, _ = tpl.Fill(Params{Checkpoint: "c", Prompt: "p", Width: 512, Height: 512, Steps: 8})
+	if steps := g["5"].(map[string]any)["inputs"].(map[string]any)["steps"]; steps != 8 {
+		t.Errorf("steps = %v, want 8", steps)
+	}
+}
+
+func TestAPlaceholderWithoutAValueOrADefaultIsSaid(t *testing.T) {
+	tpl, err := Load(`{"1":{"class_type":"X","inputs":{"a":"{{checkpoint}}","b":"size {{width}}x{{height}}"}}}`)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if _, err := tpl.Fill(Params{Width: 5}); err == nil || !strings.Contains(err.Error(), "checkpoint, height") {
+		t.Fatalf("err = %v", err)
+	}
+	g, err := tpl.Fill(Params{Checkpoint: "c", Width: 5, Height: 6})
+	if err != nil {
+		t.Fatal(err)
+	}
+	if b := g["1"].(map[string]any)["inputs"].(map[string]any)["b"]; b != "size 5x6" {
+		t.Errorf("b = %v", b)
+	}
+}
+
+func TestAWorkflowInTheUIsFormatIsRefused(t *testing.T) {
+	if _, err := Load(`{"nodes":[{"id":1}],"links":[]}`); err == nil || !strings.Contains(err.Error(), "API format") {
+		t.Fatalf("err = %v", err)
+	}
+	if _, err := Load("sdxl-turbo"); err == nil {
+		t.Fatal("an unknown built-in was loaded")
+	}
+}
+
+// fakeComfy is just enough of ComfyUI: /prompt queues, /history answers
+// after a few asks, /view serves a picture.
+type fakeComfy struct {
+	mu          sync.Mutex
+	asked       int
+	fail        string
+	refuse      bool
+	deleted     []string
+	queued      map[string]any
+	readyAt     int
+	srv         *httptest.Server
+	interrupted bool
+}
+
+func newFakeComfy(t *testing.T) *fakeComfy {
+	f := &fakeComfy{readyAt: 2}
+	f.srv = httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		f.mu.Lock()
+		defer f.mu.Unlock()
+		switch {
+		case r.Method == http.MethodPost && r.URL.Path == "/prompt":
+			if f.refuse {
+				w.WriteHeader(http.StatusBadRequest)
+				fmt.Fprint(w, `{"error":{"message":"Prompt outputs failed validation","details":""},"node_errors":{"1":{"class_type":"CheckpointLoaderSimple","errors":[{"message":"Value not in list","details":"ckpt_name: 'x' not in []"}]}}}`)
+				return
+			}
+			var b struct {
+				Prompt map[string]any `json:"prompt"`
+			}
+			_ = json.NewDecoder(r.Body).Decode(&b)
+			f.queued = b.Prompt
+			fmt.Fprint(w, `{"prompt_id":"p1","number":1,"node_errors":{}}`)
+		case r.URL.Path == "/history/p1":
+			f.asked++
+			if f.asked < f.readyAt {
+				fmt.Fprint(w, `{}`)
+				return
+			}
+			if f.fail != "" {
+				fmt.Fprintf(w, `{"p1":{"status":{"status_str":"error","completed":false,"messages":[["execution_start",{}],["execution_error",{"node_type":"KSampler","exception_message":%q}]]},"outputs":{}}}`, f.fail)
+				return
+			}
+			fmt.Fprint(w, `{"p1":{"status":{"status_str":"success","completed":true,"messages":[]},"outputs":{"7":{"images":[{"filename":"a.png","subfolder":"","type":"temp"},{"filename":"b.png","subfolder":"","type":"temp"}]}}}}`)
+		case r.URL.Path == "/view":
+			w.Header().Set("Content-Type", "image/png")
+			fmt.Fprint(w, "PNG:"+r.URL.Query().Get("filename"))
+		case r.URL.Path == "/queue" && r.Method == http.MethodPost:
+			var b struct {
+				Delete []string `json:"delete"`
+			}
+			_ = json.NewDecoder(r.Body).Decode(&b)
+			f.deleted = append(f.deleted, b.Delete...)
+		case r.URL.Path == "/queue":
+			fmt.Fprint(w, `{"queue_running":[[1,"p1",{},{},[]]],"queue_pending":[]}`)
+		case r.URL.Path == "/interrupt":
+			f.interrupted = true
+		default:
+			http.NotFound(w, r)
+		}
+	}))
+	t.Cleanup(f.srv.Close)
+	return f
+}
+
+func TestGenerateWaitsForTheWorkflowAndFetchesItsPictures(t *testing.T) {
+	f := newFakeComfy(t)
+	c := &Client{Poll: time.Millisecond}
+	images, err := c.Generate(t.Context(), f.srv.URL, "", map[string]any{"1": map[string]any{"class_type": "X"}})
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(images) != 2 || string(images[0].Data) != "PNG:a.png" || images[1].ContentType != "image/png" {
+		t.Fatalf("images = %+v", images)
+	}
+	if f.queued["1"] == nil {
+		t.Error("the workflow was not what was queued")
+	}
+}
+
+func TestAWorkflowThatFailsSaysWhy(t *testing.T) {
+	f := newFakeComfy(t)
+	f.fail = "CUDA out of memory"
+	_, err := (&Client{Poll: time.Millisecond}).Generate(t.Context(), f.srv.URL, "", map[string]any{})
+	if err == nil || !strings.Contains(err.Error(), "KSampler: CUDA out of memory") {
+		t.Fatalf("err = %v", err)
+	}
+}
+
+func TestAWorkflowComfyUIRefusesIsNotRetried(t *testing.T) {
+	f := newFakeComfy(t)
+	f.refuse = true
+	_, err := (&Client{Poll: time.Millisecond}).Generate(t.Context(), f.srv.URL, "", map[string]any{})
+	ce, ok := err.(*Error)
+	if !ok || ce.Retryable || !strings.Contains(ce.Msg, "Value not in list") {
+		t.Fatalf("err = %#v", err)
+	}
+}
+
+func TestAComfyUIThatCannotBeReachedIsRetryable(t *testing.T) {
+	f := newFakeComfy(t)
+	f.srv.Close()
+	_, err := (&Client{Poll: time.Millisecond}).Generate(t.Context(), f.srv.URL, "", map[string]any{})
+	if ce, ok := err.(*Error); !ok || !ce.Retryable {
+		t.Fatalf("err = %#v", err)
+	}
+}
+
+func TestAClientThatLeavesTakesItsWorkflowOffTheCard(t *testing.T) {
+	f := newFakeComfy(t)
+	f.readyAt = 1 << 30
+	ctx, cancel := context.WithTimeout(t.Context(), 30*time.Millisecond)
+	defer cancel()
+	_, err := (&Client{Poll: time.Millisecond}).Generate(ctx, f.srv.URL, "", map[string]any{})
+	if err == nil {
+		t.Fatal("no error")
+	}
+	f.mu.Lock()
+	defer f.mu.Unlock()
+	if len(f.deleted) != 1 || f.deleted[0] != "p1" || !f.interrupted {
+		t.Errorf("deleted %v, interrupted %v", f.deleted, f.interrupted)
+	}
+}
+
+func TestAPictureLosesTheWorkflowComfyUIWroteIntoIt(t *testing.T) {
+	chunk := func(kind, data string) []byte {
+		b := make([]byte, 0, 12+len(data))
+		b = binary.BigEndian.AppendUint32(b, uint32(len(data)))
+		b = append(b, kind...)
+		b = append(b, data...)
+		return append(b, 0, 0, 0, 0) // a CRC nobody here checks
+	}
+	png := []byte("\x89PNG\r\n\x1a\n")
+	png = append(png, chunk("IHDR", "0123456789abc")...)
+	png = append(png, chunk("tEXt", "prompt\x00{\"1\":...}")...)
+	png = append(png, chunk("IDAT", "pixels")...)
+	png = append(png, chunk("iTXt", "workflow\x00...")...)
+	png = append(png, chunk("IEND", "")...)
+	got := string(withoutPNGText(png))
+	if strings.Contains(got, "prompt") || strings.Contains(got, "workflow") {
+		t.Errorf("text left in: %q", got)
+	}
+	for _, kind := range []string{"IHDR", "IDAT", "pixels", "IEND"} {
+		if !strings.Contains(got, kind) {
+			t.Errorf("%s gone: %q", kind, got)
+		}
+	}
+	if broken := png[:30]; string(withoutPNGText(broken)) != string(broken) {
+		t.Error("a truncated PNG was changed")
+	}
+}

internal/comfy/workflows/flux-dev.json

@@ -0,0 +1,12 @@
+{
+  "1": {"class_type": "CheckpointLoaderSimple", "inputs": {"ckpt_name": "{{checkpoint}}"}},
+  "2": {"class_type": "CLIPTextEncode", "inputs": {"text": "{{prompt}}", "clip": ["1", 1]}},
+  "3": {"class_type": "CLIPTextEncode", "inputs": {"text": "{{negative_prompt}}", "clip": ["1", 1]}},
+  "8": {"class_type": "FluxGuidance", "inputs": {"guidance": "{{guidance|3.5}}", "conditioning": ["2", 0]}},
+  "4": {"class_type": "EmptySD3LatentImage", "inputs": {"width": "{{width}}", "height": "{{height}}", "batch_size": "{{n}}"}},
+  "5": {"class_type": "KSampler", "inputs": {
+    "seed": "{{seed}}", "steps": "{{steps|20}}", "cfg": 1, "sampler_name": "euler", "scheduler": "simple", "denoise": 1,
+    "model": ["1", 0], "positive": ["8", 0], "negative": ["3", 0], "latent_image": ["4", 0]}},
+  "6": {"class_type": "VAEDecode", "inputs": {"samples": ["5", 0], "vae": ["1", 2]}},
+  "7": {"class_type": "PreviewImage", "inputs": {"images": ["6", 0]}}
+}

internal/comfy/workflows/flux-schnell.json

@@ -0,0 +1,11 @@
+{
+  "1": {"class_type": "CheckpointLoaderSimple", "inputs": {"ckpt_name": "{{checkpoint}}"}},
+  "2": {"class_type": "CLIPTextEncode", "inputs": {"text": "{{prompt}}", "clip": ["1", 1]}},
+  "3": {"class_type": "CLIPTextEncode", "inputs": {"text": "{{negative_prompt}}", "clip": ["1", 1]}},
+  "4": {"class_type": "EmptySD3LatentImage", "inputs": {"width": "{{width}}", "height": "{{height}}", "batch_size": "{{n}}"}},
+  "5": {"class_type": "KSampler", "inputs": {
+    "seed": "{{seed}}", "steps": "{{steps|4}}", "cfg": 1, "sampler_name": "euler", "scheduler": "simple", "denoise": 1,
+    "model": ["1", 0], "positive": ["2", 0], "negative": ["3", 0], "latent_image": ["4", 0]}},
+  "6": {"class_type": "VAEDecode", "inputs": {"samples": ["5", 0], "vae": ["1", 2]}},
+  "7": {"class_type": "PreviewImage", "inputs": {"images": ["6", 0]}}
+}

internal/config/agent.go

@@ -18,7 +18,12 @@ import (
 //
 // An empty path loads the defaults, which with --server and --node-id on the
 // command line is enough for a smoke test on a node with no IB.
-func LoadAgent(path string) (agent.Config, error) {
+func LoadAgent(path string) (agent.Config, error) { return LoadAgentWith(path, nil) }
+
+// LoadAgentWith is LoadAgent with the cluster's settings for this node
+// (agents.default and agents.nodes.<id>, as YAML) laid over the file: the
+// file is what the agent needs to start, the settings what it runs on.
+func LoadAgentWith(path string, settings []byte) (agent.Config, error) {
 	cfg := agent.DefaultConfig()
 	k := koanf.New(".")
 
@@ -30,6 +35,11 @@ func LoadAgent(path string) (agent.Config, error) {
 			return cfg, fmt.Errorf("parsing %s: %w", path, err)
 		}
 	}
+	if len(settings) > 0 {
+		if err := k.Load(bytesProvider(settings), yaml.Parser()); err != nil {
+			return cfg, fmt.Errorf("parsing the cluster's settings for this agent: %w", err)
+		}
+	}
 
 	err := k.Load(env.Provider(".", env.Opt{
 		Prefix: "LLMGW_",

internal/config/agents.go

@@ -0,0 +1,110 @@
+package config
+
+import (
+	"fmt"
+	"slices"
+	"sort"
+)
+
+// AgentSettings are the nodes' agent settings, kept with the cluster's
+// (BACKLOG E3): everything agent.yaml can say except what an agent needs to
+// start — its id, the control planes to dial, its address and its
+// certificate. Default applies to every node, and Nodes, by node id, over
+// it; both over the node's own agent.yaml.
+type AgentSettings struct {
+	// Default is agent.yaml's keys for every node.
+	Default map[string]any `koanf:"default"`
+	// Nodes are agent.yaml's keys for one node each, by node id, over
+	// Default.
+	Nodes map[string]map[string]any `koanf:"nodes"`
+}
+
+// agentBootstrapKeys are what an agent needs before it can be told anything:
+// they stay in agent.yaml.
+var agentBootstrapKeys = []string{"node_id", "server", "servers", "hostname", "mgmt_addr", "tls"}
+
+// For is the settings for one node: Default with the node's own over it.
+// Empty when there are none.
+func (a AgentSettings) For(node string) map[string]any {
+	out := clone(stringKeys(a.Default))
+	if n, ok := a.Nodes[node]; ok {
+		merge(out, stringKeys(n))
+	}
+	return out
+}
+
+// Check refuses settings that set an agent's bootstrap keys, or that would
+// not load as an agent's configuration for some node.
+func (a AgentSettings) Check() error {
+	check := func(where string, m map[string]any) error {
+		for _, k := range agentBootstrapKeys {
+			if _, ok := m[k]; ok {
+				return fmt.Errorf("%s.%s stays in each node's agent.yaml: an agent needs it before it can be told anything", where, k)
+			}
+		}
+		return nil
+	}
+	if err := check("agents.default", stringKeys(a.Default)); err != nil {
+		return err
+	}
+	nodes := make([]string, 0, len(a.Nodes))
+	for n := range a.Nodes {
+		nodes = append(nodes, n)
+	}
+	sort.Strings(nodes)
+	for _, n := range nodes {
+		if err := check("agents.nodes."+n, stringKeys(a.Nodes[n])); err != nil {
+			return err
+		}
+	}
+	// Each node's settings as its agent would load them, over nothing: what
+	// its own file adds cannot be known here, and what is wrong in the
+	// settings alone is wrong on every node.
+	for _, n := range append([]string{""}, nodes...) {
+		b, err := Marshal(a.For(n))
+		if err != nil {
+			return err
+		}
+		if _, err := LoadAgentWith("", b); err != nil {
+			where := "agents.default"
+			if n != "" {
+				where = "agents.nodes." + n
+			}
+			return fmt.Errorf("%s: %w", where, err)
+		}
+	}
+	return nil
+}
+
+// AgentBootstrapKeys are the keys agent.yaml keeps, for saying so.
+func AgentBootstrapKeys() []string { return slices.Clone(agentBootstrapKeys) }
+
+// stringKeys turns YAML's map[any]any, wherever it hides, into the
+// map[string]any the rest of this package merges.
+func stringKeys(m map[string]any) map[string]any {
+	out := make(map[string]any, len(m))
+	for k, v := range m {
+		out[k] = normalize(v)
+	}
+	return out
+}
+
+func normalize(v any) any {
+	switch t := v.(type) {
+	case map[string]any:
+		return stringKeys(t)
+	case map[any]any:
+		m := make(map[string]any, len(t))
+		for k, vv := range t {
+			m[fmt.Sprint(k)] = normalize(vv)
+		}
+		return m
+	case []any:
+		out := make([]any, len(t))
+		for i, vv := range t {
+			out[i] = normalize(vv)
+		}
+		return out
+	}
+	return v
+}

internal/config/agents_test.go

@@ -0,0 +1,79 @@
+package config
+
+import (
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+	"time"
+)
+
+func TestANodesAgentSettingsAreTheDefaultWithItsOwnOver(t *testing.T) {
+	a := AgentSettings{
+		Default: map[string]any{"heartbeat_interval": "5s", "supervisor": map[string]any{"work_dir": "/srv/w", "x": 1}},
+		Nodes: map[string]map[string]any{
+			"cuda01": {"supervisor": map[string]any{"work_dir": "/fast/w"}, "labels": map[string]any{"gpu": "rtx6000"}},
+		},
+	}
+	got := a.For("cuda01")
+	sup := got["supervisor"].(map[string]any)
+	if sup["work_dir"] != "/fast/w" || sup["x"] != 1 || got["heartbeat_interval"] != "5s" || got["labels"] == nil {
+		t.Errorf("cuda01: %v", got)
+	}
+	if got := a.For("node01"); got["labels"] != nil || got["supervisor"].(map[string]any)["work_dir"] != "/srv/w" {
+		t.Errorf("node01: %v", got)
+	}
+	if got := (AgentSettings{}).For("x"); len(got) != 0 {
+		t.Errorf("none: %v", got)
+	}
+}
+
+func TestAgentSettingsKeepBootstrapKeysInTheFileAndMustLoad(t *testing.T) {
+	for _, a := range []AgentSettings{
+		{Default: map[string]any{"servers": []any{"x:7443"}}},
+		{Nodes: map[string]map[string]any{"n1": {"node_id": "other"}}},
+		{Nodes: map[string]map[string]any{"n1": {"tls": map[string]any{"ca": "/x"}}}},
+	} {
+		if err := a.Check(); err == nil || !strings.Contains(err.Error(), "stays in each node's agent.yaml") {
+			t.Errorf("%+v: %v", a, err)
+		}
+	}
+	// A fabric with no adapters would not load on the node it names.
+	bad := AgentSettings{Nodes: map[string]map[string]any{"n1": {"fabrics": []any{map[string]any{"name": "ib-a"}}}}}
+	if err := bad.Check(); err == nil || !strings.Contains(err.Error(), "agents.nodes.n1") || !strings.Contains(err.Error(), "no HCAs") {
+		t.Errorf("a fabric without HCAs: %v", err)
+	}
+	if err := (AgentSettings{Default: map[string]any{"heartbeat_interval": "3s"}}).Check(); err != nil {
+		t.Errorf("a plain setting: %v", err)
+	}
+}
+
+func TestTheClusterSettingsCarryAgentsAndRefuseBadOnes(t *testing.T) {
+	good, err := Parse([]byte("agents:\n  default:\n    heartbeat_interval: 3s\n  nodes:\n    cuda01:\n      labels: {gpu: rtx6000}\n"))
+	if err != nil {
+		t.Fatal(err)
+	}
+	cfg, err := FromMap(good)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if cfg.Agents.For("cuda01")["labels"] == nil {
+		t.Errorf("agents not read: %+v", cfg.Agents)
+	}
+	bad, _ := Parse([]byte("agents:\n  default:\n    mgmt_addr: 10.0.0.1\n"))
+	if _, err := FromMap(bad); err == nil {
+		t.Error("a bootstrap key in the cluster's settings was taken")
+	}
+}
+
+func TestTheSettingsAreLaidOverTheAgentsFile(t *testing.T) {
+	p := filepath.Join(t.TempDir(), "agent.yaml")
+	_ = os.WriteFile(p, []byte("node_id: n1\nserver: cp:7443\nheartbeat_interval: 9s\nlabels: {rack: r1}\n"), 0o600)
+	cfg, err := LoadAgentWith(p, []byte("heartbeat_interval: 2s\nlabels: {gpu: a100}\n"))
+	if err != nil {
+		t.Fatal(err)
+	}
+	if cfg.NodeID != "n1" || cfg.HeartbeatInterval != 2*time.Second || cfg.Labels["rack"] != "r1" || cfg.Labels["gpu"] != "a100" {
+		t.Errorf("loaded %+v", cfg)
+	}
+}

internal/config/config.go

@@ -19,6 +19,7 @@ import (
 	"gitea.mynodes.xyz/ilker/llmgw/internal/control/catalog"
 	"gitea.mynodes.xyz/ilker/llmgw/internal/control/facility"
 	"gitea.mynodes.xyz/ilker/llmgw/internal/guard"
+	"gitea.mynodes.xyz/ilker/llmgw/internal/mail"
 	"gitea.mynodes.xyz/ilker/llmgw/internal/notify"
 	"gitea.mynodes.xyz/ilker/llmgw/internal/obs"
 	"gitea.mynodes.xyz/ilker/llmgw/internal/secrets"
@@ -146,9 +147,24 @@ type Server struct {
 	// Prometheus and Alertmanager in between. Empty sends nothing.
 	Notify notify.Config `koanf:"notify"`
 
+	// Mail is where the gateway's few emails go out from: a link to set a
+	// forgotten password. Unset, none is sent and people ask an operator.
+	Mail mail.Config `koanf:"mail"`
+
+	// Serving stops and starts the whole cluster's service (llmgw stop, llmgw
+	// start).
+	Serving ServingSettings `koanf:"serving"`
+
 	// Firewall turns llmgw's firewall on every machine on or off.
 	Firewall FirewallSettings `koanf:"firewall"`
 
+	// Agents are the nodes' agent settings: agent.yaml's keys but those an
+	// agent needs to start, for every node (default) and for one
+	// (nodes.<id>). Sent to the agents as they change; an agent whose
+	// settings changed starts again on them, its engines serving
+	// throughout.
+	Agents AgentSettings `koanf:"agents"`
+
 	// FirewallNetworks are networks, besides those each machine reaches the
 	// cluster on, that llmgw's internal ports answer: a monitoring host on
 	// another subnet, nodes behind a router. Written into every machine's
@@ -707,7 +723,10 @@ type Endpoint struct {
 	Models []string `koanf:"models"`
 	// Engine names the inference engine, which decides what its metrics are
 	// called. Empty means vLLM's names.
-	Engine   string  `koanf:"engine"`
+	Engine string `koanf:"engine"`
+	// Workflow is the ComfyUI workflow for engine: comfyui, a built-in's
+	// name or its JSON; flux-schnell unless set.
+	Workflow string  `koanf:"workflow"`
 	Pool     string  `koanf:"pool"`
 	Weight   float64 `koanf:"weight"`
 	Priority int     `koanf:"priority"`
@@ -878,6 +897,9 @@ func load(b []byte) (Server, error) {
 // Validate rejects configurations that would otherwise fail confusingly at
 // request time.
 func (s Server) Validate() error {
+	if err := s.Agents.Check(); err != nil {
+		return err
+	}
 	for i, o := range s.EngineOverrides {
 		if o.GPUClass == "" && len(o.Nodes) == 0 && len(o.Labels) == 0 {
 			return fmt.Errorf("engine_overrides[%d] picks no nodes: give it a gpu_class, nodes or labels", i)
@@ -914,6 +936,9 @@ func (s Server) Validate() error {
 		return fmt.Errorf("backup_copy_to needs backup_recipients: an archive holds the certificate authority's key " +
 			"and every secret in the store, and leaves this machine only encrypted")
 	}
+	if err := s.Mail.Validate(); err != nil {
+		return err
+	}
 	if _, err := s.FirewallAllow(); err != nil {
 		return err
 	}

internal/config/serving.go

@@ -0,0 +1,11 @@
+package config
+
+// ServingSettings is whether the cluster serves at all.
+type ServingSettings struct {
+	// Stopped takes every model down and answers every request for one with
+	// 503 (system_stopped), leaving the control planes, agents and mirror
+	// running and every model's definition as it is: llmgw stop sets it,
+	// llmgw start clears it, and the models come back as their catalog
+	// entries ask. In force on every control plane within seconds.
+	Stopped bool `koanf:"stopped"`
+}

internal/control/adminapi/api.go

@@ -12,6 +12,7 @@ import (
 	"errors"
 	"fmt"
 	"io"
+	"maps"
 	"net/http"
 	"net/url"
 	"slices"
@@ -67,8 +68,16 @@ type API struct {
 	// nil without a store.
 	federation *federation.Service
 	// guard is the guardrail policies' library and attachments.
-	guard    *GuardAdmin
+	guard *GuardAdmin
+	// kb is knowledge bases; nil without a store.
+	kb *KBAdmin
+	// tenancy is organizations, keys and people; nil without a store.
+	tenancy TenancyStore
+	// train is adapter training; nil without a store and agents.
+	train    *TrainAdmin
 	settings *Settings
+	// restartAgents tells every connected agent to start again.
+	restartAgents func(reason string) []string
 	// revoke records a removed node, so its certificate is refused.
 	revoke func(ctx context.Context, nodeID string, at time.Time) error
 
@@ -111,8 +120,12 @@ func (a *API) Register(mux Mux) {
 	mux.HandleFunc("POST /api/v1/models/{name}/scale", a.scaleModel)
 	mux.HandleFunc("POST /api/v1/models/{name}/resources", a.setResources)
 	a.registerQuotas(mux)
+	a.registerKB(mux)
+	a.registerTenancy(mux)
+	a.registerTrain(mux)
 	a.registerGuard(mux)
 	a.registerSettings(mux)
+	mux.HandleFunc("POST /api/v1/agents/restart", a.restartAgentsHandler)
 	mux.HandleFunc("GET /api/v1/audit", a.auditEvents)
 	mux.HandleFunc("GET /api/v1/usage/export", a.usageExport)
 	mux.HandleFunc("POST /api/v1/notify/test", a.notifyTest)
@@ -228,6 +241,17 @@ type NodeView struct {
 	// which is not the same as having none — placement treats it that way too,
 	// or a control plane upgrade would empty a fleet.
 	Engines []string `json:"engines,omitempty"`
+	// EngineVersions are the installed engines with their versions:
+	// llamacpp → b10991. ComputeCapabilities are its GPUs' CUDA compute
+	// capabilities, 7.0 for a V100, so an engine build's support can be read
+	// against them.
+	EngineVersions map[string]string `json:"engine_versions,omitempty"`
+	// MultiNode is whether it can take part in replicas spanning nodes:
+	// "ready", "down: mlx5_1" naming the declared adapters whose ports are
+	// not active, or empty for a node on no fabric. Its own replicas run
+	// either way.
+	MultiNode           string   `json:"multi_node,omitempty"`
+	ComputeCapabilities []string `json:"compute_capabilities,omitempty"`
 	// Provision is where installing each wanted engine stands. An engine is in
 	// Engines only once its installation here is ready.
 	Provision []inventory.EngineInstall `json:"provision,omitempty"`
@@ -416,10 +440,12 @@ type ModelView struct {
 	Description string   `json:"description,omitempty"`
 	Aliases     []string `json:"aliases,omitempty"`
 	Source      string   `json:"source"`
-	Min         int      `json:"min_replicas"`
-	Max         int      `json:"max_replicas"`
-	Ready       int      `json:"ready_replicas"`
-	Total       int      `json:"total_replicas"`
+	// Role is "guard" for a guardrails judge, empty for a model clients use.
+	Role  string `json:"role,omitempty"`
+	Min   int    `json:"min_replicas"`
+	Max   int    `json:"max_replicas"`
+	Ready int    `json:"ready_replicas"`
+	Total int    `json:"total_replicas"`
 	// Desired is how many it should have now: Min, or for an on-demand
 	// model one while it is being asked for and none otherwise.
 	Desired int `json:"desired_replicas"`
@@ -518,6 +544,30 @@ type ShardView struct {
 	// node's last heartbeat; zero when not measured.
 	CPUCores    float64 `json:"cpu_cores,omitempty"`
 	MemoryBytes int64   `json:"memory_bytes,omitempty"`
+	// Compute is where the engine said it put the model as it became
+	// ready: absent when it did not say.
+	Compute *ComputeView `json:"compute,omitempty"`
+}
+
+// ComputeView is where an engine runs its model, by its own account.
+type ComputeView struct {
+	// Backend is CUDA, Vulkan, Metal, ROCm, GPU or CPU.
+	Backend string   `json:"backend"`
+	Devices []string `json:"devices,omitempty"`
+	// LayersOffloaded of LayersTotal went to the devices; absent when unsaid.
+	LayersOffloaded int `json:"layers_offloaded,omitempty"`
+	LayersTotal     int `json:"layers_total,omitempty"`
+}
+
+// Short is the compute in a few characters: "CUDA 49/49", "CPU".
+func (c *ComputeView) Short() string {
+	if c == nil {
+		return "-"
+	}
+	if c.LayersTotal > 0 && c.Backend != "CPU" {
+		return fmt.Sprintf("%s %d/%d", c.Backend, c.LayersOffloaded, c.LayersTotal)
+	}
+	return c.Backend
 }
 
 // FetchView is a download in flight.
@@ -623,6 +673,16 @@ func (a *API) Nodes() []NodeView {
 			v.Engines = append(v.Engines, name)
 		}
 		slices.Sort(v.Engines)
+		if len(n.EngineVersions) > 0 {
+			v.EngineVersions = maps.Clone(n.EngineVersions)
+		}
+		for _, g := range n.GPUs {
+			if g.ComputeCapability != "" && !slices.Contains(v.ComputeCapabilities, g.ComputeCapability) {
+				v.ComputeCapabilities = append(v.ComputeCapabilities, g.ComputeCapability)
+			}
+		}
+		slices.Sort(v.ComputeCapabilities)
+		v.MultiNode = multiNodeReadiness(n)
 		v.Provision = n.Provision
 		if ch := a.ctl.EngineChoices(); ch != nil && len(n.GPUs) > 0 {
 			chosen := ch.For(n.ID)
@@ -761,6 +821,7 @@ func (a *API) Models() []ModelView {
 			Class: m.EffectiveClass(), Owner: m.Owner, Preempted: h.Preempted, HeldFor: h.HeldFor,
 			External: externalLabel(m), Adapters: m.Adapters,
 			Canary: m.Canary, CanaryOf: a.ctl.Catalog().CanaryOf(m.Name), Eval: m.Eval,
+			Role: m.Role,
 		}
 		for _, p := range m.Profiles {
 			v.Profiles = append(v.Profiles, p.Name)
@@ -829,6 +890,10 @@ func (a *API) Replicas(model string) []ReplicaView {
 					sv.Fetch.Percent = &p
 				}
 			}
+			if c := s.Compute; c != nil {
+				sv.Compute = &ComputeView{Backend: c.Backend, Devices: c.Devices,
+					LayersOffloaded: c.Offloaded, LayersTotal: c.Total}
+			}
 			if l := loads[s.NodeID]; l != nil {
 				if use, ok := l.Shards[s.ID]; ok {
 					sv.CPUCores, sv.MemoryBytes = use.CPUCores, use.RSSBytes
@@ -1473,6 +1538,11 @@ type RestartRequest struct {
 	// InPlace stops each replica before starting its replacement, rather
 	// than starting the replacement beside it first.
 	InPlace bool `json:"in_place,omitempty"`
+	// MaxSurge is how many of a model's replicas are replaced beside
+	// themselves at once, MaxUnavailable how many are stopped first at once;
+	// both zero is one at a time, replaced beside itself.
+	MaxSurge       int `json:"max_surge,omitempty"`
+	MaxUnavailable int `json:"max_unavailable,omitempty"`
 }
 
 // RestartResult lists the replicas queued.
@@ -1496,11 +1566,12 @@ func (a *API) restart(w http.ResponseWriter, r *http.Request) {
 		writeError(w, http.StatusBadRequest, "nothing ready matches; name replicas, models, or all")
 		return
 	}
-	restart := a.ctl.Reconciler().Restart
-	if req.InPlace {
-		restart = a.ctl.Reconciler().RestartInPlace
+	if req.MaxSurge < 0 || req.MaxUnavailable < 0 {
+		writeError(w, http.StatusBadRequest, "max_surge and max_unavailable are counts of replicas, not negative")
+		return
 	}
-	if err := restart(ids...); err != nil {
+	pace := reconciler.RestartPace{MaxSurge: req.MaxSurge, MaxUnavailable: req.MaxUnavailable}
+	if err := a.ctl.Reconciler().RestartPaced(pace, req.InPlace, ids...); err != nil {
 		writeError(w, http.StatusNotFound, err.Error())
 		return
 	}
@@ -1545,6 +1616,7 @@ type RequestView struct {
 	GuardAction   string `json:"guard_action,omitempty"`
 	GuardPolicy   string `json:"guard_policy,omitempty"`
 	GuardCategory string `json:"guard_category,omitempty"`
+	GuardRule     string `json:"guard_rule,omitempty"`
 	GuardMs       int64  `json:"guard_ms,omitempty"`
 	// What the client said about it: OpenAI's user field, and its tags.
 	EndUser string            `json:"end_user,omitempty"`
@@ -1573,7 +1645,7 @@ func (a *API) request(w http.ResponseWriter, r *http.Request) {
 		Model: u.Model, ServedModel: u.ServedModel, Replica: u.Replica, Status: u.Status,
 		LatencyMs: u.LatencyMs, QueueMs: u.QueueMs, TTFTMs: u.TTFTMs,
 		PromptTokens: u.PromptTokens, CompletionTokens: u.CompletionTokens, CostMicros: u.CostMicros,
-		GuardAction: u.GuardAction, GuardPolicy: u.GuardPolicy, GuardCategory: u.GuardCategory, GuardMs: u.GuardMs,
+		GuardAction: u.GuardAction, GuardPolicy: u.GuardPolicy, GuardCategory: u.GuardCategory, GuardRule: u.GuardRule, GuardMs: u.GuardMs,
 		EndUser: u.EndUser, Tags: u.Tags, CacheHit: u.CacheHit,
 	})
 }
@@ -1897,32 +1969,9 @@ func (a *API) setAdapters(w http.ResponseWriter, r *http.Request) {
 		writeError(w, http.StatusBadRequest, "no model "+name)
 		return
 	}
-	base, baseErr := gguf.KindFromURL(r.Context(), safedial.Client(safedial.ControlPlane, 30*time.Second), m.Source)
-	have := map[string]bool{}
-	for _, ad := range m.Adapters {
-		have[ad.Source] = true
-	}
-	// The file is read only where llama.cpp will load it: a vLLM adapter is
-	// a PEFT directory, which has no header to read, and vLLM itself refuses
-	// one that does not fit the model when it is loaded.
-	ggufs := slices.ContainsFunc(m.Profiles, func(p catalog.Profile) bool { return p.Engine == "llamacpp" })
-	for _, ad := range req.Adapters {
-		if have[ad.Source] || !ggufs {
-			continue // already checked when it was attached
-		}
-		h, err := gguf.KindFromURL(r.Context(), safedial.Client(safedial.ControlPlane, 30*time.Second), ad.Source)
-		switch {
-		case err != nil:
-			writeError(w, http.StatusBadRequest, fmt.Sprintf("adapter %s: %v", ad.Name, err))
-			return
-		case !h.IsAdapter():
-			writeError(w, http.StatusBadRequest, fmt.Sprintf("adapter %s: %s is a model, not a LoRA adapter", ad.Name, ad.Source))
-			return
-		case baseErr == nil && h.Architecture != base.Architecture:
-			writeError(w, http.StatusBadRequest, fmt.Sprintf("adapter %s is for %s models and %s is %s",
-				ad.Name, h.Architecture, m.Name, base.Architecture))
-			return
-		}
+	if err := checkAdapters(r.Context(), m, req.Adapters); err != nil {
+		writeError(w, http.StatusBadRequest, err.Error())
+		return
 	}
 	out, err := a.models.SetAdapters(r.Context(), m.Name, req.Adapters, actorOf(r))
 	if err != nil {
@@ -1975,3 +2024,65 @@ func (a *API) SetNodeEngine(ctx context.Context, node, engine, build, by string)
 	}
 	return ch.Set(ctx, node, engine, build, by)
 }
+
+// multiNodeReadiness is whether a node's declared fabrics are up: every
+// adapter they name with an active port. Empty for a node on no fabric.
+func multiNodeReadiness(n inventory.Node) string {
+	if len(n.Fabrics) == 0 {
+		return ""
+	}
+	if len(n.IBPorts) == 0 {
+		return "unknown: no port reported yet"
+	}
+	var down []string
+	for _, f := range n.Fabrics {
+		for _, hca := range f.HCAs {
+			up := false
+			for _, p := range n.IBPorts {
+				if p.HCA == hca && p.Active() {
+					up = true
+				}
+			}
+			if !up && !slices.Contains(down, hca) {
+				down = append(down, hca)
+			}
+		}
+	}
+	if len(down) > 0 {
+		return "down: " + strings.Join(down, ", ")
+	}
+	return "ready"
+}
+
+// checkAdapters reads each adapter a model does not have yet, where
+// llama.cpp will load it, to refuse one that is not a LoRA adapter for the
+// model's architecture: such an adapter is a replica that fails to start,
+// found minutes later on a node.
+func checkAdapters(ctx context.Context, m catalog.Model, adapters []catalog.Adapter) error {
+	// A vLLM adapter is a PEFT directory, which has no header to read, and
+	// vLLM itself refuses one that does not fit the model when it is loaded.
+	if !slices.ContainsFunc(m.Profiles, func(p catalog.Profile) bool { return p.Engine == "llamacpp" }) {
+		return nil
+	}
+	client := safedial.Client(safedial.ControlPlane, 30*time.Second)
+	base, baseErr := gguf.KindFromURL(ctx, client, m.Source)
+	have := map[string]bool{}
+	for _, ad := range m.Adapters {
+		have[ad.Source] = true
+	}
+	for _, ad := range adapters {
+		if have[ad.Source] {
+			continue // checked when it was attached
+		}
+		h, err := gguf.KindFromURL(ctx, client, ad.Source)
+		switch {
+		case err != nil:
+			return fmt.Errorf("adapter %s: %v", ad.Name, err)
+		case !h.IsAdapter():
+			return fmt.Errorf("adapter %s: %s is a model, not a LoRA adapter", ad.Name, ad.Source)
+		case baseErr == nil && h.Architecture != base.Architecture:
+			return fmt.Errorf("adapter %s is for %s models and %s is %s", ad.Name, h.Architecture, m.Name, base.Architecture)
+		}
+	}
+	return nil
+}

internal/control/adminapi/guard.go

@@ -28,6 +28,10 @@ type GuardAdmin struct {
 	Library    func() guard.Config
 	Store      GuardStore
 	Invalidate func()
+	// JudgeErrors are each judge's last failure to give a verdict.
+	JudgeErrors func() map[string]guard.JudgeError
+	// Ask has a judge read a text, past the cache.
+	Ask func(ctx context.Context, model, text, answer string) (guard.JudgeAnswer, error)
 }
 
 // SetGuard enables the guard endpoints.
@@ -35,6 +39,8 @@ func (a *API) SetGuard(g *GuardAdmin) { a.guard = g }
 
 // GuardView is the library and where its policies are attached.
 type GuardView struct {
+	// Enabled is whether the guardrails check anything (guard.enabled).
+	Enabled     bool                  `json:"enabled"`
 	Policies    []string              `json:"policies"`
 	Default     []string              `json:"default"`
 	Attachments []GuardAttachmentView `json:"attachments"`
@@ -89,7 +95,10 @@ type GuardEventView struct {
 	Action    string    `json:"action"`
 	Policy    string    `json:"policy"`
 	Category  string    `json:"category,omitempty"`
-	GuardMs   int64     `json:"guard_ms"`
+	// Rule is what decided it within the category: a pattern's name, or a
+	// judge's verdict, "S6,Specialized Advice".
+	Rule    string `json:"rule,omitempty"`
+	GuardMs int64  `json:"guard_ms"`
 }
 
 // policyView sums one policy up.
@@ -198,6 +207,8 @@ type GuardAttachmentRequest struct {
 }
 
 func (a *API) registerGuard(mux Mux) {
+	mux.HandleFunc("GET /api/v1/guard/judges", a.listJudges)
+	mux.HandleFunc("POST /api/v1/guard/judges/ask", a.askJudge)
 	mux.HandleFunc("GET /api/v1/guard", func(w http.ResponseWriter, r *http.Request) {
 		if a.guard == nil {
 			writeError(w, http.StatusNotImplemented, "no guard policies: define them under guard in the cluster's settings (llmgw ctl config edit)")
@@ -208,7 +219,7 @@ func (a *API) registerGuard(mux Mux) {
 			writeError(w, http.StatusInternalServerError, err.Error())
 			return
 		}
-		v := GuardView{Default: a.guard.Library().Default, Attachments: []GuardAttachmentView{}}
+		v := GuardView{Enabled: a.guard.Library().On(), Default: a.guard.Library().Default, Attachments: []GuardAttachmentView{}}
 		var names map[string]string
 		if a.quotas != nil {
 			names, _ = a.scopeNames(r.Context())
@@ -269,7 +280,7 @@ func (a *API) registerGuard(mux Mux) {
 		slices.SortStableFunc(v.Counts, func(a, b GuardCountView) int { return int(b.Count - a.Count) })
 		for _, e := range events {
 			v.Events = append(v.Events, GuardEventView{At: e.At, RequestID: e.RequestID, Project: project(e.ProjectID),
-				Model: e.Model, Status: e.Status, Action: e.Action, Policy: e.Policy, Category: e.Category, GuardMs: e.GuardMs})
+				Model: e.Model, Status: e.Status, Action: e.Action, Policy: e.Policy, Category: e.Category, Rule: e.Rule, GuardMs: e.GuardMs})
 		}
 		writeJSON(w, http.StatusOK, v)
 	})

internal/control/adminapi/judges.go

@@ -0,0 +1,200 @@
+package adminapi
+
+import (
+	"encoding/json"
+	"fmt"
+	"net/http"
+	"slices"
+	"sort"
+	"strings"
+	"time"
+
+	"gitea.mynodes.xyz/ilker/llmgw/internal/control/reconciler"
+	"gitea.mynodes.xyz/ilker/llmgw/internal/guard"
+)
+
+// JudgeView is one judge model the guardrail policies ask, and whether it
+// can answer them.
+type JudgeView struct {
+	Model string `json:"model"`
+	// Policies are those that ask it; Closed those that refuse what it
+	// cannot read (fail: closed).
+	Policies []string `json:"policies"`
+	Closed   []string `json:"closed,omitempty"`
+	// Then are the judges its policies ask after it, when it gives no
+	// verdict: a CPU judge behind a GPU one.
+	Then []string `json:"then,omitempty"`
+	// InCatalog is whether a model of that name is in the catalog, and
+	// Judge whether it is there as a judge (role: guard).
+	InCatalog bool `json:"in_catalog"`
+	Judge     bool `json:"judge"`
+	Ready     int  `json:"ready_replicas"`
+	Wanted    int  `json:"wanted_replicas"`
+	// LastError is its last failure to give a verdict, if any.
+	LastError   string     `json:"last_error,omitempty"`
+	LastErrorAt *time.Time `json:"last_error_at,omitempty"`
+	// Problem says, in words, what is wrong and what to do; empty when
+	// nothing is.
+	Problem string `json:"problem,omitempty"`
+}
+
+// judgeErrorFresh is how recent a judge's failure must be to count as a
+// problem now.
+const judgeErrorFresh = 5 * time.Minute
+
+// Judges are the judge models the policies in force ask, each with what is
+// wrong with it, if anything.
+func (a *API) Judges() []JudgeView {
+	if a.guard == nil || a.guard.Library == nil {
+		return nil
+	}
+	lib := a.guard.Library()
+	if !lib.On() {
+		// Off: no judge is asked, so none is missing.
+		return nil
+	}
+	byModel := map[string]*JudgeView{}
+	for name, p := range lib.Policies {
+		judges := p.Judge.Judges()
+		for i, model := range judges {
+			v := byModel[model]
+			if v == nil {
+				v = &JudgeView{Model: model}
+				byModel[model] = v
+			}
+			if !slices.Contains(v.Policies, name) {
+				v.Policies = append(v.Policies, name)
+			}
+			if p.Fail == "closed" && p.Mode != "shadow" && !slices.Contains(v.Closed, name) {
+				v.Closed = append(v.Closed, name)
+			}
+			for _, after := range judges[i+1:] {
+				if !slices.Contains(v.Then, after) {
+					v.Then = append(v.Then, after)
+				}
+			}
+		}
+	}
+	ready := map[string]int{}
+	for _, rep := range a.ctl.Reconciler().Replicas() {
+		if rep.State == reconciler.Ready {
+			ready[rep.Model]++
+		}
+	}
+	var errs map[string]guard.JudgeError
+	if a.guard.JudgeErrors != nil {
+		errs = a.guard.JudgeErrors()
+	}
+	out := make([]JudgeView, 0, len(byModel))
+	for _, v := range byModel {
+		sort.Strings(v.Policies)
+		sort.Strings(v.Closed)
+		if m, ok := a.ctl.Catalog().Resolve(v.Model); ok && m.Name == v.Model {
+			v.InCatalog, v.Judge = true, m.IsGuard()
+			v.Wanted, v.Ready = m.Scaling.Min, ready[m.Name]
+		}
+		if e, ok := errs[v.Model]; ok {
+			at := e.At
+			v.LastError, v.LastErrorAt = e.Err, &at
+		}
+		v.Problem = judgeProblem(*v)
+		out = append(out, *v)
+	}
+	slices.SortFunc(out, func(x, y JudgeView) int {
+		switch {
+		case x.Model < y.Model:
+			return -1
+		case x.Model > y.Model:
+			return 1
+		}
+		return 0
+	})
+	return out
+}
+
+func judgeProblem(v JudgeView) string {
+	consequence := "requests its policies judge are served unjudged"
+	switch {
+	case len(v.Then) > 0:
+		consequence = "its policies ask " + strings.Join(v.Then, ", else ") + " instead"
+	case len(v.Closed) > 0:
+		consequence = "requests its policies judge are refused (guard_unavailable)"
+	}
+	switch {
+	case !v.InCatalog:
+		return fmt.Sprintf("not in the catalog, so %s: llmgw ctl guard judge enable", consequence)
+	case !v.Judge:
+		return fmt.Sprintf("in the catalog as a model clients use, not as a judge (role: guard), so %s: "+
+			"llmgw ctl guard judge enable --replace", consequence)
+	case v.Ready == 0:
+		return fmt.Sprintf("no replica ready, so %s: llmgw ctl replicas --model %s", consequence, v.Model)
+	case v.LastErrorAt != nil && time.Since(*v.LastErrorAt) < judgeErrorFresh:
+		return fmt.Sprintf("gave no verdict %s ago (%s)", time.Since(*v.LastErrorAt).Round(time.Second), v.LastError)
+	}
+	return ""
+}
+
+func (a *API) listJudges(w http.ResponseWriter, _ *http.Request) {
+	if a.guard == nil {
+		writeError(w, http.StatusNotImplemented, "no guard policies: define them under guard in the cluster's settings (llmgw ctl config edit)")
+		return
+	}
+	writeJSON(w, http.StatusOK, a.Judges())
+}
+
+// JudgeAskRequest is a text to have a judge read.
+type JudgeAskRequest struct {
+	// Model is the judge; empty is the one the policies ask, when they ask one.
+	Model string `json:"model,omitempty"`
+	Text  string `json:"text"`
+	// Answer, when given, is read as the assistant's reply to Text.
+	Answer string `json:"answer,omitempty"`
+}
+
+// JudgeAskView is what the judge answered.
+type JudgeAskView struct {
+	Model string `json:"model"`
+	// Verdict is safe, unsafe or controversial, as the gateway reads the
+	// answer; a policy's judge.controversial says what controversial counts as.
+	Verdict    string   `json:"verdict"`
+	Categories []string `json:"categories,omitempty"`
+	// Raw is the judge's answer as it gave it.
+	Raw    string `json:"raw"`
+	TookMs int64  `json:"took_ms"`
+}
+
+func (a *API) askJudge(w http.ResponseWriter, r *http.Request) {
+	if a.guard == nil || a.guard.Ask == nil {
+		writeError(w, http.StatusNotImplemented, "no guard policies: define them under guard in the cluster's settings (llmgw ctl config edit)")
+		return
+	}
+	var req JudgeAskRequest
+	if err := json.NewDecoder(r.Body).Decode(&req); err != nil || strings.TrimSpace(req.Text) == "" {
+		writeError(w, http.StatusBadRequest, "give the text for the judge to read")
+		return
+	}
+	if req.Model == "" {
+		judges := a.Judges()
+		if len(judges) == 0 {
+			writeError(w, http.StatusBadRequest, "no policy asks a judge: give the model")
+			return
+		}
+		req.Model = judges[0].Model
+	}
+	ans, err := a.guard.Ask(r.Context(), req.Model, req.Text, req.Answer)
+	v := JudgeAskView{Model: req.Model, Raw: ans.Raw, Categories: ans.Categories, TookMs: ans.Took.Milliseconds()}
+	if err != nil {
+		writeError(w, http.StatusBadGateway, fmt.Sprintf("the judge %s gave no verdict: %v (it answered %q)", req.Model, err, ans.Raw))
+		return
+	}
+	switch {
+	case ans.Unsafe:
+		v.Verdict = "unsafe"
+	case ans.Controversial:
+		// Qwen3Guard's middle verdict: what it counts as is the policy's.
+		v.Verdict = "controversial"
+	default:
+		v.Verdict = "safe"
+	}
+	writeJSON(w, http.StatusOK, v)
+}

internal/control/adminapi/judges_test.go

@@ -0,0 +1,125 @@
+package adminapi
+
+import (
+	"context"
+	"strings"
+	"testing"
+	"time"
+
+	"gitea.mynodes.xyz/ilker/llmgw/internal/control/catalog"
+	"gitea.mynodes.xyz/ilker/llmgw/internal/control/inventory"
+	"gitea.mynodes.xyz/ilker/llmgw/internal/guard"
+)
+
+func TestEachJudgeSaysWhatIsWrongWithIt(t *testing.T) {
+	models := append(testModels(),
+		// Served from the mirror page before it knew better: a model
+		// clients use, under the judge's name.
+		catalog.Model{Name: "shield", Source: "/s", Scaling: catalog.Scaling{Min: 1, Max: 1},
+			Profiles: []catalog.Profile{{Name: "cpu", Engine: "llamacpp"}}},
+		catalog.Model{Name: "llama-guard-3-1b", Source: "/g", Role: catalog.RoleGuard,
+			Guard: &catalog.GuardModel{Format: "llamaguard3"}, Scaling: catalog.Scaling{Min: 2, Max: 2},
+			Profiles: []catalog.Profile{{Name: "cpu", Engine: "llamacpp"}}})
+	a, _ := setup(t, models)
+	lib := guard.Config{Default: []string{"baseline"}, Policies: map[string]guard.Policy{
+		"baseline": {Fail: "closed", Judge: &guard.JudgePolicy{Model: "llama-guard-3-1b", Input: "block"}},
+		"strict":   {Fail: "closed", Judge: &guard.JudgePolicy{Model: "missing-judge", Input: "block"}},
+		"shield":   {Fail: "open", Judge: &guard.JudgePolicy{Model: "shield", Input: "flag"}},
+	}}
+	a.SetGuard(&GuardAdmin{Library: func() guard.Config { return lib }})
+
+	var js []JudgeView
+	if code := call(t, a, "GET", "/api/v1/guard/judges", "", &js); code != 200 || len(js) != 3 {
+		t.Fatalf("%d %+v", code, js)
+	}
+	by := map[string]JudgeView{}
+	for _, j := range js {
+		by[j.Model] = j
+	}
+	if j := by["missing-judge"]; j.InCatalog || !strings.Contains(j.Problem, "not in the catalog") ||
+		!strings.Contains(j.Problem, "refused") {
+		t.Errorf("missing: %+v", j)
+	}
+	if j := by["shield"]; !j.InCatalog || j.Judge || !strings.Contains(j.Problem, "--replace") ||
+		!strings.Contains(j.Problem, "unjudged") {
+		t.Errorf("not a judge: %+v", j)
+	}
+	if j := by["llama-guard-3-1b"]; !j.Judge || j.Wanted != 2 || !strings.Contains(j.Problem, "no replica ready") {
+		t.Errorf("not running: %+v", j)
+	}
+
+	var s Summary
+	call(t, a, "GET", "/api/v1/summary", "", &s)
+	if len(s.Judges) != 3 {
+		t.Errorf("the summary does not carry the judges' problems: %+v", s.Judges)
+	}
+
+	// A judge that failed lately says so, once it is otherwise fine.
+	at := time.Now()
+	v := JudgeView{Model: "x", InCatalog: true, Judge: true, Ready: 1, LastError: "deadline exceeded", LastErrorAt: &at}
+	if p := judgeProblem(v); !strings.Contains(p, "deadline exceeded") {
+		t.Errorf("a recent failure: %q", p)
+	}
+	old := at.Add(-time.Hour)
+	v.LastErrorAt = &old
+	if p := judgeProblem(v); p != "" {
+		t.Errorf("an old failure still counts: %q", p)
+	}
+}
+
+func TestAJudgeIsAskedAndItsWordsShown(t *testing.T) {
+	a, _ := setup(t, testModels())
+	lib := guard.Config{Policies: map[string]guard.Policy{
+		"baseline": {Judge: &guard.JudgePolicy{Model: "llama-guard-3-1b", Input: "block"}}}}
+	var asked string
+	a.SetGuard(&GuardAdmin{Library: func() guard.Config { return lib },
+		Ask: func(_ context.Context, model, text, _ string) (guard.JudgeAnswer, error) {
+			asked = model + "|" + text
+			return guard.JudgeAnswer{Model: model, Raw: "unsafe\nS6", Unsafe: true,
+				Categories: []string{"S6", "Specialized Advice"}}, nil
+		}})
+	var v JudgeAskView
+	if code := call(t, a, "POST", "/api/v1/guard/judges/ask", `{"text":"how do I sort a list"}`, &v); code != 200 {
+		t.Fatalf("%d", code)
+	}
+	if asked != "llama-guard-3-1b|how do I sort a list" || v.Verdict != "unsafe" || v.Raw != "unsafe\nS6" ||
+		len(v.Categories) != 2 {
+		t.Errorf("asked %q, got %+v", asked, v)
+	}
+	if code := call(t, a, "POST", "/api/v1/guard/judges/ask", `{"text":"  "}`, nil); code != 400 {
+		t.Errorf("no text: %d", code)
+	}
+}
+
+func TestANodeSaysItsEnginesVersionsAndCapabilities(t *testing.T) {
+	n := gpuNode("volta", 2)
+	n.EngineVersions = map[string]string{"llamacpp": "b10991", "vllm": "0.10.2"}
+	for i := range n.GPUs {
+		n.GPUs[i].ComputeCapability = "7.0"
+	}
+	a, _ := setup(t, testModels(), n)
+	var nodes []NodeView
+	call(t, a, "GET", "/api/v1/nodes", "", &nodes)
+	if len(nodes) != 1 || nodes[0].EngineVersions["vllm"] != "0.10.2" ||
+		len(nodes[0].ComputeCapabilities) != 1 || nodes[0].ComputeCapabilities[0] != "7.0" {
+		t.Errorf("%+v", nodes)
+	}
+}
+
+func TestANodeSaysWhetherItsFabricsAreUp(t *testing.T) {
+	n := gpuNode("ib1", 2, "ib-a")
+	if got := multiNodeReadiness(n); !strings.HasPrefix(got, "unknown") {
+		t.Errorf("no port yet: %q", got)
+	}
+	n.IBPorts = []inventory.IBPort{{HCA: "mlx5_0", State: "DOWN"}}
+	if got := multiNodeReadiness(n); got != "down: mlx5_0" {
+		t.Errorf("down: %q", got)
+	}
+	n.IBPorts[0].State = "ACTIVE"
+	if got := multiNodeReadiness(n); got != "ready" {
+		t.Errorf("up: %q", got)
+	}
+	if got := multiNodeReadiness(gpuNode("plain", 1)); got != "" {
+		t.Errorf("no fabric: %q", got)
+	}
+}