3 Commits

Author	SHA1	Message	Date
josedario87	5bd115e197	perf: incremental transcript updates via WebSocket, eliminate polling ... - Replace 1500ms polling with fs.watch recursive (reliable on Windows) - Enrich WS broadcast with newContent delta + session metadata - Client appends incrementally instead of 2 sequential HTTP requests - Pre-initialize Tauri HTTP plugin at module load to avoid dynamic import overhead - Per-file debounce timers (150ms) instead of single shared timer - Size-based validation for safe incremental appends with HTTP fallback	2026-02-24 11:01:54 -06:00
josedario87	65303df96a	feat: Samsung lock screen face widget, voice assistant services, PiP mode and gitignore installers ... Add Samsung proprietary Face Widget (lock screen/AOD) with terminal status display. Add voice interaction services (AgentVoiceInteractionService, RecognitionService) for digital assistant registration. Add PiP mode with voice/expand actions. Add session-state proxy, voice transcript routes, window controls component. Ignore installers/ directory.	2026-02-23 20:52:11 -06:00
josedario87	e1aa8b1bdb	feat: integrate Tauri v2 with Android widget and voice assistant ... - Add Tauri v2 shell (Cargo, tauri.conf.json, capabilities, plugins) - Migrate all fetch() calls to apiFetch() for Tauri-aware HTTP - Migrate WebSocket endpoints to resolveEndpoints() for dynamic URLs - Add ServerConfigDialog for remote server URL configuration - Add tauri.ts lib with isTauri detection, apiFetch wrapper, plugin helpers - Add server-config Pinia store with persistence via plugin-store - Conditional PWA (disabled in Tauri builds) - Android: home screen transcript widget (last 5 messages, 30s refresh) - Android: voice command / share activity (SpeechRecognizer + WebSocket) - Android: signed release APK with auto-copy to installers/ - Remove stale frontend/src-tauri directory	2026-02-23 15:33:43 -06:00