LAURA-agent/LAURA-MCP-Server

LAURA MCP Server - Clean Deployment

A comprehensive Model Context Protocol (MCP) server for the LAURA voice assistant system.

Project Structure

LAURA-MCP-Clean/
├── core/                 # Core MCP server components
├── TTS/                  # Text-to-Speech server (ElevenLabs)
├── voice_injection/      # Voice input and injection system
├── bridge/              # Mac hub bridge for iOS connectivity
├── memory/              # Memory management and persistence
├── agents/              # Specialized agent modules
├── hydration/           # Hydration tracking system
├── email_system/        # Email management and integration
├── llm_integrations/    # LLM adapter modules
├── config/              # Configuration files
├── scripts/             # Launch and management scripts
├── logs/                # System logs
├── chat_logs/           # Conversation history
├── audio_cache/         # TTS audio cache
└── working_memory/      # Session-specific context

Quick Start

1. Setup Virtual Environment

python3 -m venv venv
source venv/bin/activate
pip install -r config/requirements.txt

2. Configure API Keys

Create core/secret.py with your API keys:

ANTHROPIC_API_KEY = "your-key"
OPENAI_API_KEY = "your-key"
ELEVENLABS_API_KEY = "your-key"

3. Launch Services

Start MCP Server

./scripts/start_mcp_server.sh

Start TTS Server

./TTS/start_tts_server.sh

Start Voice Input

./voice_injection/start_voice_input.sh

Start Mac Hub Bridge (for iOS)

python bridge/mac_http_bridge.py

Or launch everything at once:

./scripts/launch_laura_full.sh

Core Components

MCP Server (core/mcp_server.py)

• Main server implementation using FastMCP
• Handles device registration and session management
• Coordinates all subsystems

TTS Server (TTS/tts_server.py)

• ElevenLabs integration for text-to-speech
• Audio streaming and caching
• Personal voice support

Voice Injection (voice_injection/quick_voice.py)

• Captures voice input
• Injects into Claude Code or terminal
• Push-to-talk and wake word support

Mac Hub Bridge (bridge/mac_http_bridge.py)

• HTTP bridge for iOS devices
• Translates MCP protocol to REST API
• Manages device sessions

Memory System (memory/)

• Persistent conversation context
• Topic mapping and timeline
• Working memory for active sessions

Agent System (agents/)

• Specialized agents for different tasks
• Calendar, contacts, email agents
• Claude Code integration

Configuration

Environment Variables

• ANTHROPIC_API_KEY: Claude API access
• OPENAI_API_KEY: GPT model access
• ELEVENLABS_API_KEY: TTS service

Port Configuration

• MCP Server: 8765
• TTS Server: 9090
• Mac Hub Bridge: 8080

Dependencies

See config/requirements.txt for full list. Key dependencies:

• fastmcp
• anthropic
• openai
• elevenlabs
• fastapi
• uvicorn
• websockets

System Requirements

• macOS 10.15+
• Python 3.10+
• 4GB RAM minimum
• Network access for API calls

Daemon Setup (Optional)

For automatic startup:

./scripts/setup_mcp_daemon.sh

Troubleshooting

Check Service Status

curl http://localhost:8765/health

View Logs

tail -f logs/mcp_server.log

Reset Memory

rm -rf working_memory/*

License

Private project - All rights reserved

Support

For issues or questions, check the documentation in the docs/ folder.