feat: local offline voice models (sherpa-onnx) with model manager and wake word (v2.1.0)

- electron/voice: model catalog (Whisper base/small/turbo, SenseVoice, Kokoro v1.0,
  Piper fr), streaming tar.bz2 downloader with progress, sherpa-onnx worker running in
  an Electron utilityProcess (transcribe / synthesize / status / unload), IPC + bridge.
- Renderer: 'local' providers for STT and TTS, Settings → Modèles locaux (download,
  progress, delete, activate), speaker selection, hands-free wake word with a tolerant
  matcher (accents, punctuation, edit distance) and attention window after a bare wake word.
- Packaging: native addon and worker unpacked from the asar; release 2.1.0.

Validated on Linux: Kokoro (fr) → Whisper base round trip, Piper (fr) → Whisper round trip.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017Wn5VX9HNbJ7N54hR24u9Y

electron/main.ts

@@ -8,6 +8,8 @@ import { registerTelemetryIpc } from './ipc/telemetry';
 import { getSharedDirectory, registerFilesIpc } from './ipc/files';
 import { createMainWindow, getMainWindow, getWindowMode, setWindowMode, toggleWindowVisibility } from './window';
 import { getWebhookStatus, startWebhookServer, stopWebhookServer } from './webhook';
+import { registerVoiceIpc } from './voice/ipc';
+import { stopEngine } from './voice/engine';
 
 // Audio playback must never be blocked behind a user gesture (TTS starts on incoming events).
 app.commandLine.appendSwitch('autoplay-policy', 'no-user-gesture-required');
@@ -149,6 +151,7 @@ app.whenReady().then(async () => {
   registerTelemetryIpc();
   registerFilesIpc();
   registerCoreIpc();
+  registerVoiceIpc();
   createMainWindow();
   createTray();
   registerShortcuts();
@@ -162,6 +165,7 @@ app.whenReady().then(async () => {
 app.on('will-quit', () => {
   globalShortcut.unregisterAll();
   abortAllStreams();
+  stopEngine();
   void stopWebhookServer();
 });