## Summary
Extract runtime-neutral provider definitions from `@proj-airi/stage-ui`
into a new `@proj-airi/provider-inference` package that runs in both
Node.js and browser without Vue/Pinia dependencies.
## What changed
- **New package** `packages/provider-inference/` — owns
`ProviderDefinition`, `ProviderRegistry`, validators, and all provider
modules (cloud + local)
- **Provider modules relocated** — ~45 providers moved from
`stage-ui/src/libs/providers/providers/` to
`provider-inference/src/providers/cloud/` and
`provider-inference/src/providers/local/`
- **Type decoupling** — `ProviderContext.t` changed from
`ComposerTranslation` (Vue) to generic `ProviderTranslator`, removing
the Vue dependency from core types
- **stage-ui types slimmed** — `types.ts` now re-exports from
`provider-inference` and only adds the Vue-specific `ProviderViews`
extension
- **validators/run.ts** — becomes a pass-through re-export
- **Registry** — portable `createProviderRegistry()` with deterministic
ordering and duplicate-id detection
## Verification
```bash
pnpm -F @proj-airi/provider-inference typecheck
pnpm -F @proj-airi/provider-inference test:node
pnpm -F @proj-airi/provider-inference test:browser
pnpm -F @proj-airi/provider-inference build
pnpm -F @proj-airi/stage-ui typecheck
pnpm -F @proj-airi/stage-web typecheck
pnpm -F @proj-airi/stage-tamagotchi typecheck
pnpm lint
```
## Why
Provider definitions were tightly coupled to Vue (`ComposerTranslation`,
`Component`), preventing reuse in non-Vue runtimes (Node.js services,
tests, future Electron backend). This extraction creates a clean
boundary: runtime-neutral definitions live in `provider-inference`,
while Vue/Pinia-specific concerns stay in `stage-ui`.
## Summary
Adds a self-contained better-auth plugin
(`server/apps/api/src/libs/auth-plugins/steam.ts`) implementing Steam
OpenID 2.0 sign-in, account linking, and callback verification via "dumb
mode".
Steam's web login is OpenID 2.0, not OAuth2/OIDC, so it cannot be
registered as a `socialProviders` entry, and better-auth has no plugin
hook for extending its OAuth2 endpoints with a non-OAuth2 protocol. The
plugin therefore adds the endpoints Steam's protocol needs: `POST
/sign-in/steam`, `POST /link/steam`, and `GET /steam/callback`.
- Callback verification uses OpenID "dumb mode"
(`openid.mode=check_authentication`): one extra round trip to Steam
instead of managing RSA association state.
- New sign-ups get a placeholder `<steamid64>@steam.placeholder.local`
with `emailVerified: true`, mirroring Apple Sign In's
`<sub>@apple.placeholder.local`.
- The plugin's request/query schemas use Zod; a `// NOTICE:` documents
that better-auth's OpenAPI generator is Zod-native. Steam verification
uses `ofetch`.
- Wires Steam into `apps/ui-server-auth` sign-in and profile "Connected
accounts", plus the shared `OAuthProvider` / `defaultSignInProviders` in
`packages/stage-ui`.
- Linking routes through `/link/steam` via the client's `$fetch`;
unlinking needs no special-casing (`/unlink-account` already takes a
free-form `providerId`).
No Steam Web API key is required for this browser-based flow.
We intentionally do not depend on community Steam packages (e.g.
`better-auth-steam`) or the still-open upstream draft
([better-auth#4877](https://github.com/better-auth/better-auth/pull/4877)).
Steam never returns an email, and we need sign-up that does not ask the
user for one plus first-class account linking; the available options
either require an email at sign-in, lack linking, or are abandoned /
blocked — shipping a small in-tree plugin is the safer auth dependency
for this requirement.
## Test plan
- [x] `pnpm exec vitest run
server/apps/api/src/libs/auth-plugins/steam.test.ts` — 6/6 passing
- [x] `pnpm -F @proj-airi/ui-server-auth exec vitest run` — 32/32
passing
- [x] `pnpm -F @proj-airi/stage-ui exec vitest run
src/libs/steam-auth-client.test.ts
src/composables/use-linked-accounts.test.ts` — 5/5 passing
- [x] `pnpm -F @proj-airi/api-server typecheck`
- [x] `pnpm -F @proj-airi/ui-server-auth typecheck`
- [x] `pnpm -F @proj-airi/stage-ui typecheck`
## Follow-ups
- Desktop Steam ticket sign-in (top of this stack): silent startup
ticket exchange for Steam builds; the server resolves or creates the
AIRI user for the verified SteamID before issuing an OIDC code.
- Steam persona name/avatar via `GetPlayerSummaries` inside the plugin,
if display names beyond `Steam User <id>` are wanted.
---------
Co-authored-by: Cursor <cursoragent@cursor.com>
Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>
- Removed specific catalog references for '@unocss/core', 'es-toolkit', 'nanoid', and 'valibot' in pnpm-workspace.yaml and various service package.json files.
- Updated versions for '@vueuse/core', 'es-toolkit', 'nanoid', 'reka-ui', and 'valibot' to their latest compatible versions.
- Cleaned up unused catalog entries in pnpm-workspace.yaml.
Depends on #1487
## Summary
- Introduce `packages/core-agent` — a new pure-runtime package that
extracts zero-Vue/Pinia algorithmic logic from `@proj-airi/stage-ui`,
following a "compatible facade + core sinking" strategy
- Migrate shared chat types, LLM streaming types, chat hook registry,
session message merge logic, context registry algorithm, and LLM service
utilities into `core-agent`
- `stage-ui` files are preserved as re-exports / thin wrappers — all
external imports, store IDs, and public APIs remain unchanged
## Motivation
`stage-ui` currently mixes pure agent runtime logic (algorithms, type
definitions, stateless utilities) with Vue/Pinia state management and
browser-specific adapters. This coupling makes it hard to:
- Test agent logic in isolation
- Reuse agent algorithms outside of Vue contexts (e.g., server-side,
CLI, other frameworks)
- Reason about the boundary between "what the agent does" vs "how the UI
manages state"
---------
Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>
Co-authored-by-agent: Unknown <unknown@example.com>
## Summary
Address maintainer feedback from #1622 ([@nekomeowww's
comment](https://github.com/moeru-ai/airi/pull/1622#discussion_r2342632282)).
The project already uses `async-mutex` in other packages
(`stage-tamagotchi`, `electron-screen-capture`), and it's in the
workspace catalog. This replaces the custom `AsyncMutex` implementation
with the shared dependency.
- Replace `AsyncMutex.run()` → `Mutex.runExclusive()`
- Replace `AsyncMutex.reset()` → `Mutex.cancel()`
- Remove `async-mutex.ts` and its unit tests
- Remove `AsyncMutex` from inference barrel exports
- Add `async-mutex` as direct dependency of `@proj-airi/stage-ui`
**Regarding `@moeru/eventa` suggestion**
([comment](https://github.com/moeru-ai/airi/pull/1622#discussion_r2342631692)):
The `waitForMessage` pattern is a thin request-response abstraction over
`postMessage`, where the worker is driven by `@huggingface/transformers`
internally. Eventa's transport-agnostic RPC is better suited for
bidirectional channels (Electron IPC, WebSocket) than for this one-way
"post and wait" pattern. No change for now.
## Test plan
- [x] `pnpm exec vitest run packages/stage-ui/src/libs/inference/` — 11
tests pass (4 removed with custom impl)
- [x] `pnpm -F @proj-airi/stage-ui typecheck` — no errors
- [x] `pnpm lint:fix` — no new errors in changed files
---------
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>
## Problem
The WebGPU inference pipeline had several structural issues:
1. **No unified protocol** — Kokoro TTS, Whisper ASR, and Background
Removal workers each used their own ad-hoc message formats. Adding a new
model meant reinventing worker communication from scratch.
2. **Infrastructure existed but was disconnected** —
`GPUResourceCoordinator`, `LoadQueue`, `InferenceWorkerManager`, and
`protocol.ts` were all implemented but had zero consumers. The adapters
duplicated the same lifecycle/timeout/mutex patterns independently.
3. **Performance gaps** — Kokoro only offered fp32 on WebGPU (no fp16),
Whisper warm-up compiled shaders for 187.5s of dummy audio, audio
transfer went through unnecessary WAV blob encode/decode, and
`listVoices` reloaded the model every time.
4. **Silent failures** — Whisper worker's `generate()` had no try-catch;
errors were swallowed and the main thread waited until timeout.
5. **No graceful degradation** — Whisper and Background Removal workers
hardcoded `device: 'webgpu'` with no WASM fallback.
6. **No observability** — Only Kokoro had performance tracing. No
adapter reported status to `useInferenceStatus`. No cache management UI
existed.
7. **Dead code accumulation** — Old `KokoroWorkerManager` (232 lines),
legacy Whisper message types, and scattered duplicate constants.
## Changes
### Phase 0 — Critical Performance & Bugs
- Add `fp16-webgpu` dtype for Kokoro TTS (~2x inference speed on
supported GPUs)
- Fix Whisper warm-up tensor from `[1, 128, 3000]` → `[1, 128, 1]`
(minimal shader compilation)
- Fix Whisper worker silent error bug (add try-catch to `generate()` and
`load()`)
### Phase 1 — Data Transfer & Caching
- Switch Kokoro audio to Float32Array transferable (skip WAV blob encode
in worker, lightweight WAV encode on main thread)
- Cache `listVoices` results (skip redundant model reload when adapter
state is `ready`)
- Normalize progress reporting to 0-100 across all adapters,
differentiate `warmup` phase
### Phase 2 — Protocol Unification & Infrastructure
- Migrate all 3 workers + 3 adapters to unified `protocol.ts` message
types (`load-model`, `run-inference`, `model-ready`, `inference-result`,
`progress`, `error`)
- Wire `GPUResourceCoordinator` into all adapters (VRAM allocation
tracking, LRU ordering, memory pressure events)
- Wire `LoadQueue` into all adapters (priority-based sequential model
loading: TTS=10 > ASR=5 > BG_REMOVAL=1)
- Add `coordinator.ts` global singleton for GPU coordinator + load queue
- Add WebGPU detection + WASM fallback in Whisper and Background Removal
workers
### Phase 3 — Error Recovery & Observability
- Add restart logic with exponential backoff to Whisper adapter
(matching Kokoro's existing pattern)
- Integrate `classifyError()` (OOM / DEVICE_LOST / TIMEOUT
classification) in Whisper adapter
- Extend `defaultPerfTracer` to Whisper `transcribe()` and Background
Removal `processImage()`
- Wire `useInferenceStatus` into all 3 adapters (downloading → ready →
terminated lifecycle)
### Phase 4 — Tests
- Add unit tests for `AsyncMutex` (4 tests), `LoadQueue` (4 tests),
`GPUResourceCoordinator` (7 tests) — all 15 passing
### Phase 5 — Cleanup & Features
- Delete old `KokoroWorkerManager` (232 lines, zero consumers)
- Delete orphaned `libs/workers/types.ts` (old Whisper message types)
- Clean up `workers/kokoro/types.ts` (remove legacy message types, keep
domain types)
- Create centralized `constants.ts` (MODEL_IDS, MODEL_NAMES, TIMEOUTS,
MAX_RESTARTS)
- Remove hardcoded WebGPU check from background-removal devtools pages
(worker auto-detects)
- Add `useModelPreload` composable for generic idle-time preloading
- Add `useInferencePreload` composable that reads provider config and
preloads configured local models
- Wire preloading into both `stage-web` and `stage-tamagotchi` App.vue
(Kokoro TTS preloads 3s after init)
- Add `ModelCacheManager.vue` settings component (cache size display,
per-model status, clear cache)
- Document GPU Device isolation architecture in protocol.ts
## After This PR
- All inference workers speak the same protocol → adding a new model
adapter is straightforward
- GPU memory is tracked across all models with automatic pressure
warnings at 80%/95% of VRAM budget
- Models load sequentially via priority queue → no bandwidth/VRAM
contention
- Workers auto-detect WebGPU and fall back to WASM → works on browsers
without WebGPU
- Kokoro TTS preloads during idle time → "instant" first use for
configured users
- All adapters auto-restart on worker crashes (max 3 attempts,
exponential backoff)
- 15 unit tests cover core infrastructure (mutex, queue, coordinator)
- Zero dead code remains in the inference pipeline
## Test Plan
- [x] `pnpm exec vitest run packages/stage-ui/src/libs/inference/` — 15
tests pass
- [x] `pnpm -F @proj-airi/stage-ui exec tsc --noEmit` — no TypeScript
errors
- [x] `pnpm lint:fix` — no lint errors in changed files
- [ ] Manual: verify Kokoro TTS works with fp16-webgpu on a supported
browser
- [ ] Manual: verify Whisper ASR loads and transcribes correctly
- [ ] Manual: verify Background Removal works in devtools page
- [ ] Manual: verify preloading triggers in console (`[Preload] Loading
kokoro-...`)
---------
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>
RainbowBird