The official CLI marks meta/muse-spark-* as reasoning models but ships no
effort levels for them and sends no effort parameter, so the thinking level
stays off in pi. Document that so the locked footer is not mistaken for a
provider bug.
Refs #69
(cherry picked from commit 263310635a89e35cf0b3da527474555c9687903e)
Every host start awaited a full catalog request before the provider
registered, costing one HTTPS round-trip (up to the discovery timeout on a
hanging connection) even when a cache written seconds earlier was on disk.
Register the cached catalog immediately and run the live refresh in the
background; the live result re-registers the provider when it arrives. A
first start without a cache still awaits the live catalog. The background
refresh is aborted on session_shutdown so print mode does not wait for it.
Closes#63
(cherry picked from commit 142191f9420fe90460cdc3cbae70e26d0c000860)
Declare the fixture's pi types inline so npm run typecheck passes without
the optional peer packages, matching the existing fixture convention.
(cherry picked from commit 0ba5827c58a0f6ccf54387454ae38d93bbead3d9)
pi routes the main chat through the registered provider, but sibling
extensions that call streamSimple from @earendil-works/pi-ai/compat with
the active Command Code model resolve model.api through the compat
api-registry, which only knows built-in APIs. On plain pi that failed
with "No API provider registered for api: commandcode-custom".
Register commandcode-custom there and delegate to the transport router.
The registry resolves no credentials for extension providers, so fall
back to the configured Command Code key when the caller passes none.
Closes#68
(cherry picked from commit 7e9659e672771c6a9223b95938e50fb2a051a0a7)
Verify display pricing against the official pricing page and the live
provider model list:
- add claude-fable-5-1 and deepseek/deepseek-v4-flash-fast
- remove minimax/minimax-m3-free and minimax/minimax-m2.7-free, which the
provider no longer lists
- end the claude-sonnet-5 introductory window; the listed rate is unchanged
- google/gemini-3.7-flash returns to list price after the promotion ended
Add claude-fable-5-1, deepseek/deepseek-v4-flash-fast, and
tencent/hy4-preview with their reasoning efforts, add moonshotai/Kimi-K3
efforts, and drop minimax/minimax-m3-free, which the CLI no longer lists.
The model catalog tests hard-coded specific model ids, the reasoning model
count, and output limits from command-code@1.32.2. Every upstream catalog
sync broke them, which made the daily catalog sync workflow fail before it
could open its PR.
Assert structural invariants over the generated catalog instead: image
models resolve to text+image, every effort entry has a reasoning flag,
reasoning without efforts yields an empty level map, output limits are
positive integers and clamp to the context length.
Closes#67
(cherry picked from commit 458e3a57892bb625a78fb5ded092627ba3cf2867)
Maps the three Muse Spark models to [minimal, low, medium, high, xhigh]
so pi-ai can surface graded thinking levels instead of off-only.
These ids are already MODEL_REASONING=true; without MODEL_EFFORTS,
thinkingMetadataForModel() returns an empty map and the footer stays
:off with no cycle via /thinking or Shift+Tab.
Upstream command-code@1.32.2 lists these models as Efforts: — (none);
this commit intentionally adds a pi-side policy mapping aligned with
the opencode/opencode-go provider catalogs (which ship
muse-spark@ minimal..xhigh). The catalog header notes it is normally
generated by sync:commandcode-catalog — happy to move this to a
separate override layer if maintainers prefer.
Add an agent skill that walks through a full model catalog refresh: drift
detection, catalog sync, manually reviewed pricing updates, fixture
refresh, and test updates. Its helper scripts run on Windows and Linux:
one snapshots the live model-id list into the test fixture, the other
regenerates the pricing fixture from MODEL_COSTS through prettier so
format:check stays green. Typecheck now covers the skill scripts.
execFile cannot spawn npm's .cmd shim directly on Windows, so the catalog
sync and drift check failed with spawn npm ENOENT. Route npm invocations
through the shell with argument quoting on Windows and keep direct
execFile calls elsewhere.
Add the free minimax/minimax-m3-free and minimax/minimax-m2.7-free
promotional variants (free through September 5, 2026) and
tencent/hy4-preview, drop the retired stealth/ox-alpha, and refresh the
model-id and pricing snapshots to the current 62-model catalog verified
on 2026-08-28.
Qwen/Qwen3.8-Flash lists $0.16 input, $0.47 output, and $0.016 cache read
per million tokens, and z-ai/glm-5.3-flash lists $0.15, $0.50, and $0.03;
the official pricing page documents no cache-write rate for either model.
Synchronize the static model catalog with command-code@1.36.0, which adds
Qwen/Qwen3.8-Flash (reasoning efforts low, medium, xhigh) and
z-ai/glm-5.3-flash (low, high, max), the free minimax/minimax-m3-free
model, and the glm-5.3-flash output limit, and drops the retired
stealth/ox-alpha.
Align the developer-message fix with the Provider API transport router.
The fix stays scoped to the legacy /alpha/generate converter, and the
OMP advisory compat test now forces the upgrade_required fallback so the
advisory is asserted on the generate request body.
Patrick Wozniak