Files
paseo/packages/server
Mohamed Boudra d3aee8c2df fix(server): hide Pi custom messages marked display: false
pi-goal-x sends its `<pi_goal_continuation …/>` checkpoint trigger as a custom
message with `display: false`: the model reads it, and Pi's TUI never renders
it. Paseo ignored the flag and emitted every custom message as assistant text,
which the timeline projection then glued onto the previous answer. The live
and history paths now read custom text through one function that respects the
flag. Extension adapters still see hidden messages, so structured mappings are
unchanged; pi-subagents' hidden `subagent-notify` text is dropped the same way.

Also removes the unused `mapCustomMessage` history hook left over from the
Oh My Pi removal.
2026-09-26 19:56:48 +02:00
..
2026-04-23 21:44:48 +07:00

Voice Assistant

A voice-controlled terminal assistant that runs as a single local service.

Quick Start

# Install dependencies
npm install

# Copy environment variables
cp .env.example .env
# Edit .env and add your API keys (OpenAI, Deepgram)

# Run development servers
npm run dev

# Open browser to http://localhost:5173

Architecture

  • Express Server (port 3000) - Serves API and built UI in production
  • Vite Dev Server (port 5173) - Hot-reload React UI in development
  • WebSocket (/ws) - Real-time bidirectional communication
  • Agent - STT → LLM → TTS pipeline with terminal control
  • Daemon - tmux-based terminal management (in-process)

Development

# Run both servers (recommended)
npm run dev

# Or run separately:
npm run dev:server  # Express on port 3000
npm run dev:ui      # Vite on port 5173

# Type checking
npm run typecheck

# Build for production
npm run build

# Start production server
npm start

Project Status

✅ Completed (Phases 1-2):

  • Package setup and configuration
  • Express server with WebSocket
  • React UI with Vite
  • WebSocket client with ping/pong testing

⏳ In Progress (Phase 3):

  • Terminal control (tmux integration)

📋 Planned (Phases 4-9):

  • LLM integration (OpenAI GPT-4)
  • Agent orchestrator
  • Speech-to-Text (Deepgram)
  • Text-to-Speech (OpenAI)
  • Audio streaming
  • UI polish

See IMPLEMENTATION_PLAN.md for complete details.

Environment Variables

OPENAI_API_KEY=your-openai-key-here      # GPT-4 and TTS
DEEPGRAM_API_KEY=your-deepgram-key-here  # Streaming STT
STT_MODEL=whisper-1        # Optional: override to gpt-4o-transcribe, etc.
STT_CONFIDENCE_THRESHOLD=-3.0  # Optional: reject low-confidence clips
STT_DEBUG_AUDIO_DIR=.stt-debug # Optional: persist raw dictation audio for debugging
PASEO_HOME=~/.paseo        # Runtime state directory (agents/, etc.)
PASEO_LISTEN=127.0.0.1:6767  # Listen address (host:port or /path/to/socket)

PASEO_HOME defaults to ~/.paseo and isolates runtime artifacts like agents/. PASEO_LISTEN controls the daemon listen address. For blue/green testing you can run a parallel server without touching production state:

PASEO_HOME=~/.paseo-blue PASEO_LISTEN=127.0.0.1:7777 npm run dev

Tech Stack

  • Server: Express, TypeScript, ws (WebSocket)
  • Client: React 18, Vite, TypeScript
  • Terminal: tmux (via child_process)
  • AI: OpenAI (LLM + TTS), Deepgram (STT)

Testing

Currently manual testing via:

  1. Start servers: npm run dev
  2. Open http://localhost:5173
  3. Test WebSocket connection (green status indicator)
  4. Click "Send Ping" button to test communication

More testing guidance as features are implemented.

License

MIT