Every provider in agent/models.yml is a LAN address, so a workstation that
leaves the house has no model at all. Add `openwebui (franksdank.cloud)`:
the already-public OpenWebUI vhost as an OpenAI-compatible provider that
proxies the same engines, with frank-dragoon and frank-zealot rows and the
WAN row as its own compaction target. LAN rows, modelRoles and the speech
runner stay untouched — the door is purely additive.
The key is command-backed (`apiKey: "!cat ~/.config/secrets/openwebui_api_key"`,
mode 600) instead of inline: a literal sk-... in a tracked file is a key in
every agent context on every box that pulls this repo, and the pre-commit
shape gate would block it anyway. Rotate = rewrite the file + restart omp,
because a `!cmd` value is cached for the process lifetime.
Compat blocks copy their LAN twins because the proxy was probed, not assumed
(2026-10-21): streaming SSE carries reasoning_content and tool_calls deltas
plus final usage, image_url content parts reach both vision towers, and
chat_template_kwargs is forwarded — proven on frank-zealot, where
enable_thinking false yields zero reasoning and true a real block.
frank-archon gets no WAN row: OpenWebUI answers 400 "Model not found"
because its picker never published it, so that path is a fleet change.
Verified end-to-end from a scratch HOME whose LAN baseUrls point at a dead
port, so any answer can only have come through the door (text and a real
tool round-trip); `omp models` still lists all four chat providers and the
tts runner, so the file validates. bats 137/137, kb-lint clean.
Docs: omp-docs/core/wan-model-access.md.
The entry rewrote every effort before the wire: `minimal -> low`, `high -> xhigh`,
`max -> xhigh`, under a comment claiming flash-server rejects anything outside
{xhigh, medium, low}. The engine says otherwise, in its own 400:
reasoning_effort 'maximum' unsupported; use minimal|low|medium|high|xhigh|max|none
and the harness passes an unmapped level through verbatim
(`compat.reasoningEffortMap?.[effort] ?? effort`, read in the pinned `dist/cli.js`). So the
fold protected nothing and silently changed what the caller asked for. The levels are not
cosmetic either — thinking length on one prompt at `max_tokens 2048`: `low` 766 reasoning
tokens, `xhigh` 614, `max` 1049 with the block closed at 1024.
The map is identity. Two things came out of capturing the wire (scratch HOME + a logging
proxy forwarding :7799 to :7770) that reading the config could not have told me:
- `:off` does not switch thinking off. The default disable mode sends the lowest effort,
so `off` left as `reasoning_effort: "minimal"` — and `minimal` still thinks, a canned
26-token block on every request. `reasoningDisableMode: none-effort` makes the harness
send `"none"`, which the engine honours as a real switch: `reasoning_tokens: 0`, answer
only, no thinking block rendered. Verified after the change, not before it.
- `:max` still arrives as `xhigh`: the harness clamps the requested effort to its own
ladder for this model before the map applies. Recorded in the comment as the engine-side
half of the contract rather than papered over — nothing in models.yml forces it.
Also pinned `top_k: 20` (accepted; the comment claimed the wire matched the server sampler
while `top_k` was never sent, leaving the recipe one env edit away from drifting), and
restored the pre-callsign model ids `frank-medium` / `frank-large` / `frank-small` — the
2026-10-08 rename had replaced the old id inside its own explanation, leaving "renamed
from frank-dragoon" three times and destroying the record needed to resolve an old id.
Verified: full suite green (0 failures), `kb-lint` 0 violations, YAML parses, and the four
efforts `off/high/medium/max` captured on the wire as `none/high/medium/xhigh` with
`top_k: 20` on every request. Refs #62
Uncommitted machine/session edits (live config, models, docs drift) were
blocking the merge of origin/main. Committed as-is to preserve them; the
merge that follows resolves overlapping prose files toward upstream
canonical, with local variants recoverable from this commit. Pre-merge
copies also in .AGENTTRASH/pre-merge-2026-10-18/.
frank-medium -> frank-archon (dankserver), frank-large -> frank-dragoon
(dankclaw-gpu), frank-small -> frank-zealot (dankclaw-3090). The Dankclaw
service/unit names keep the old frank-large spelling; only the OMP model ids
move, per the comments at each entry.
Windows refit to the served context minus the compaction headroom:
archon 196608 -> 200000 ctx, 32768 -> 16384 maxTokens
dragoon 200192 -> 217088 ctx, 57344 -> 45056 maxTokens, input +image
zealot 262144 -> 200704 ctx, 32768 -> 16384 maxTokens
All three now carry compactionModel frank-zealot, reasoningContentField
reasoning_content, and Qwen-recommended sampling in compat.extraBody
(merged last, so it beats the global -1 sampler settings).
- task.agentModelOverrides.scout: "@smol" (frank-small, xhigh thinking)
- task.maxConcurrency: 2 (semaphore on ACTIVE subagents; total
instantiations uncapped)
- compaction.remoteEndpoint -> frank-medium /v1/chat/completions
(192.168.0.235:7770), methodOrder remote-first; compactionModel:
frank-medium on every model so the wire model id is always valid
- models.yml header: single source of truth (pi-era import retired)
Verified: omp config get on all keys; omp -p startup probe passed;
live POST in the remote-compaction wire format answered.
Also carries previously-uncommitted edits already in the working file:
maxTokens 262144 -> 65536 (HALOGEN_MAX_TOKENS_CAP) with the KV-pool
admission note, and removal of the flm (dankclaw-npu) provider block.
Live check: halogen /health reports context=262144, so 200000 is inside
the served window; 200000 + 65536 = 265536 against a 401408-position pool.
halogen-flash-server has no server-side sampler defaults (temperature absent
or 0 = greedy), so the endpoint decoded greedy for OMP, whose global
sampling settings default to -1 (not sent).
compat.extraBody merges these top-level into every request body after the
compat policy, so they win:
temperature=1.0 top_p=0.95 top_k=20 min_p=0.0
presence_penalty=0.0 repetition_penalty=1.0 preserve_thinking=true
Verified against the live dankclaw endpoint: preserve_thinking is honoured
(350 vs 75 prompt tokens for a replayed reasoning turn), and the exact
body shape returns 200.
Frank