feat(voice): first-class voice input and local TTS across web, desktop, and mobile (#2018)

Complete rebuild of voice input on a server-authoritative streaming
architecture, replacing the legacy Web Speech / whole-blob / WASM engines
and the dead voice-agent layer (~4k lines removed).

Speech-to-text (dictation):
- Client streams 16 kHz mono PCM16 chunks over /api/dictation/ws with
  seq/ack ordering; buffered audio is retained and replayed on reconnect
- Server transcribes and streams live partial transcripts back;
  segments auto-commit every ~15s with silence suppression and adaptive
  finalization timeouts
- Local provider (default, zero config): sherpa-onnx models in a forked
  worker process — auto-download with progress, staged extraction with
  verification, corrupt-model auto-recovery, idle shutdown after 5 min
- Model catalog with settings picker (accuracy/speed ratings, sizes,
  download/delete): Parakeet TDT v2 (English) and v3 (25 European
  languages, auto-detected), Whisper base and tiny (multilingual, light)
- OpenAI-compatible provider for any Whisper endpoint
- Composer overlay with live transcript, volume meter, timer, and
  cancel / insert / insert-and-send actions; failed transcriptions keep
  their audio for retry or accepting the partial text as-is
- Configurable keyboard shortcut (default mod+alt+v) toggles dictation;
  Enter confirms and Escape cancels while recording
- Overlay is pixel-aligned with the composer (measured footer height,
  matching paddings/typography/gaps) — no layout shift when toggling

Text-to-speech:
- Local Kokoro provider (English, 11 voices) synthesized in the same
  worker via /api/dictation/tts/speak, managed by the shared model
  pipeline; sentence-pipelined playback keeps time-to-first-audio at
  ~1 sentence regardless of message length, and stop cancels in-flight
  synthesis
- Sanitizer keeps inline-code content (strips backticks only), reads
  interword slashes aloud, and removes only absolute file paths

Settings:
- Voice page unified: a single read-aloud toggle owns all playback
  options (the confusing "Enable Voice Mode" is gone); a new "Enable
  voice input" toggle (default on, persisted to settings.json) hides
  the composer mic entirely when disabled

Mobile and transport:
- iOS/Android microphone permissions added (dictation was previously
  impossible on mobile)
- Fixed Android WebSocket upgrades: the Capacitor WebView origin
  (https://localhost) was missing from the packaged-client allowlist,
  403-ing every WS connection — root cause of the old mobile SSE lock,
  which is now removed for all transports

Security and conventions:
- All HTTP routes sit behind the global /api auth gate; the WS upgrade
  explicitly validates the UI session and origin, with oc_url_token
  narrowly allowlisted and covered by tests; the dictation socket mints
  a fresh URL token before connecting
- Routes register before the generic OpenCode proxy; the client goes
  through runtimeFetch/getRuntimeUrlResolver, and runtime switches
  reset the dictation socket
- VS Code deliberately reports dictation as unavailable (no server
  process in that runtime)

CI: workflow Node bumped 20 -> 22 to match the repo engines and fix
better-sqlite3 installs broken by node-gyp@latest on Node 20.

New dependency: sherpa-onnx-node (prebuilt N-API; macOS/Linux x64+arm64,
Windows x64 — Windows-on-ARM falls back to the OpenAI-compatible provider)

packages/ui/src/lib/i18n/messages/es.settings.ts

@@ -1485,16 +1485,33 @@ export const settingsDict = {
   "settings.notifications.page.toast.backgroundDisabled": "Notificaciones de fondo deshabilitadas",
   "settings.notifications.page.testNotification.title": "Notificación de prueba",
   "settings.notifications.page.testNotification.body": "Esta es una notificación de prueba de OpenChamber.",
-  "settings.voice.page.section.voiceSetup": "Configuración de voz",
   "settings.voice.page.section.speechRecognition": "Reconocimiento de voz",
+  "settings.voice.page.field.enableVoiceInput": "Habilitar entrada de voz",
+  "settings.voice.page.field.enableVoiceInputAria": "Habilitar entrada de voz (dictado)",
   "settings.voice.page.section.playbackAndSummary": "Reproducción",
-  "settings.voice.page.field.enableVoiceModeAria": "Habilitar modo de voz",
-  "settings.voice.page.field.enableVoiceMode": "Habilitar modo de voz",
   "settings.voice.page.field.provider": "Proveedor",
   "settings.voice.page.provider.browser": "Navegador",
   "settings.voice.page.provider.custom": "Personalizado",
   "settings.voice.page.provider.say": "Decir",
   "settings.voice.page.provider.server": "Servidor",
+  "settings.voice.page.provider.local": "Local",
+  "settings.voice.page.tooltip.sttLocal": "Transcripción local en el servidor de OpenChamber. Los modelos se descargan automáticamente; no se necesita clave de API.",
+  "settings.voice.page.tooltip.localTts": "Síntesis local en el servidor de OpenChamber (Kokoro, inglés). El modelo se descarga automáticamente; no se necesita clave de API.",
+  "settings.voice.page.stt.model.parakeetV2": "Parakeet v2 (inglés)",
+  "settings.voice.page.stt.model.parakeetV3": "Parakeet v3 (25 idiomas europeos)",
+  "settings.voice.page.stt.model.whisperBase": "Whisper base (multilingüe)",
+  "settings.voice.page.stt.model.whisperTiny": "Whisper tiny (multilingüe)",
+  "settings.voice.page.stt.badge.bestForEnglish": "Mejor para inglés",
+  "settings.voice.page.stt.badge.bestForMultilingual": "Mejor multilingüe",
+  "settings.voice.page.stt.meta.accuracy": "Precisión",
+  "settings.voice.page.stt.meta.speed": "Velocidad",
+  "settings.voice.page.stt.modelInstalled": "Modelo instalado",
+  "settings.voice.page.stt.modelDownloading": "Descargando modelo...",
+  "settings.voice.page.stt.modelDownloadingProgress": "Descargando modelo... {percent}%",
+  "settings.voice.page.stt.modelNotInstalled": "Modelo no descargado",
+  "settings.voice.page.stt.modelDownload": "Descargar",
+  "settings.voice.page.stt.modelDelete": "Eliminar modelo",
+  "settings.voice.page.stt.modelRetry": "Reintentar",
   "settings.voice.page.provider.wasm": "Local",
   "settings.voice.page.stt.wasmModel": "Modelo Whisper",
   "settings.voice.page.stt.wasmLoaded": "Modelo cargado",