## What
Closes the three structural gaps in AIRI's product analytics: the signup
surface had zero instrumentation, the payment funnel had no terminator
in PostHog, and SPA route changes emitted no pageviews. Also adds
semantic events for character cards, desktop-only features, and
data-maintenance actions.
## User paths
- User signs up / logs in / verifies email / resets password / links
OAuth / deletes account → each step now emits a PostHog event from
`apps/ui-server-auth` (previously fully uninstrumented), with
`identify()` wired on session load so anonymous funnel events merge into
the user person.
- User pays via Stripe → webhook writes `product_events` as before, and
the product-events service now forwards `payment_completed` (plus signup
and subscription lifecycle facts) to PostHog via posthog-node
`captureImmediate`, keyed by the Better Auth user id → the
`checkout_started → payment_completed` funnel closes end-to-end.
Per-request LLM/TTS volume is explicitly not forwarded.
- User navigates between routes in any surface (web / desktop / pocket /
docs) → `$pageview` + `$pageleave` fire per route change via the
posthog-js `defaults: '2025-05-24'` preset in the shared
`posthog.config.ts`.
- User creates / imports / duplicates / edits a ccv3 card, switches
stage background, runs destructive data actions (export / import / clear
chats, reset providers, wipe app data), or uses desktop differentiators
(Spotlight send, widget windows, in-app updater, MCP server management,
pairing QR) → dedicated low-cardinality events.
## Notable decisions
- Server forwarding defaults on: `POSTHOG_PROJECT_KEY` defaults to the
shared browser-safe phc_* project key; set it to an empty string to
disable. Postgres `product_events` remains the source of truth.
- Cross-surface events (`oauth_callback_failed`, account lifecycle)
share one stage vocabulary exported from stage-ui so the two emitters
cannot drift silently.
- Events captured right before full-page navigation use `sendBeacon` so
they survive the redirect (checkout, OAuth consent handoff, login
redirect).
- Removed dead wrappers (`trackSignup`, `trackFirstModelSelected`,
`trackModelChanged`) that duplicated live event streams under second
names.
## How tested
- `pnpm -F @proj-airi/server exec vitest run
src/services/domain/product-events.test.ts` — 6 passed, covering the
forwarding allowlist, the `user_signed_up → signup_completed` mapping,
non-forwarded per-request actions, and a throwing sink not failing the
webhook path nor losing the DB row.
- stage-ui suites (`use-analytics`, `use-linked-accounts`, exports
contract) — 22 passed, including new account/card/data/desktop event
assertions.
- Real transport smoke: posthog-node `captureImmediate` against
`us.i.posthog.com` with the production key resolved in 1380ms (one
`server_forwarding_smoke_test` event left in the project; filter by
event name).
- Browser-tested pageviews: `VITE_ENABLE_POSTHOG=true` dev build, two
`history.pushState` route changes each produced a `$pageview` with
`$pathname`, `navigation_type: pushState`, previous-page dwell time, and
the `surface` super property; batched POST to `us.i.posthog.com/e/`
returned 200. Note: posthog-js drops events from automated browsers
(`navigator.webdriver`) by default — the verification session bypassed
the bot filter locally; production config is untouched.
- Typecheck and lint pass for server, stage-ui, stage-pages, stage-web,
stage-tamagotchi, ui-server-auth.
Full verification record:
`apps/server/docs/ai-context/verifications/posthog-forwarding-and-pageview.md`
## Follow-ups (not in this PR)
- Bot channel usage stats (Discord / Telegram) once they route through
server-runtime counters.
- Main-process desktop events (tray menu, global shortcut fire) need
renderer relay plumbing.
- Confirm `payment_completed` arrives in PostHog after the first real
Stripe payment post-deploy.
https://claude.ai/code/session_01Q1yGavkQ1P41YhTWE4XKex
- 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.
- Added a new PostHog client for capturing server-side business events such as Stripe webhooks and subscription state changes.
- Implemented various tracking functions for pricing funnel steps, character creation, and chat session starts.
- Enhanced the flux meter tests to handle partial charges and report unbilled flux correctly.
- Updated the CharacterDialog and Flux settings pages to track user interactions with analytics events.
- Introduced a mechanism to identify users on PostHog based on authentication state to ensure accurate funnel tracking.
- Added necessary dependencies for PostHog integration in the project.
## Description
<!-- Please insert your description here and especially provide info
about the "what" this PR is solving -->
`vue-router` is producing a warning about duplicate route `/devtools` on
`stage-web`. Removing `stage-web/src/pages/devtools/index.vue` resolves
the warning.
There is a duplicate route introduced in
`stage-pages/src/pages/devtools/index.vue`. The file in `stage-web` is
removed in favor of the one in `stage-pages`.
related commit
[`44ec5e4`](https://github.com/moeru-ai/airi/commit/44ec5e416c04ab6fe1d39f01f67330703078439d)
## Linked Issues
<!-- Optional, if you have any -->
## Additional Context
<!-- e.g. is there anything you'd like reviewers to focus on? -->
---------
Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>
Co-authored-by: Neko <neko@ayaka.moe>
## 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>
Neko Ayaka