claude-voice-mcp

Claude Voice MCP

This project is an implementation of an MCP server that extends Claude Desktop with voice conversation capabilities, focusing initially on text-to-speech (TTS) conversion. It supports real-time communication via WebSocket, multiple languages, and includes a simple test client for verification. The project is part of the CogentEcho.ai ecosystem, integrating with various strategic, tactical, and foundational layers.

Features

• MCP server for Claude Desktop
• Text-to-Speech conversion
• WebSocket-based communication
• Multiple language support
• Test client included

Development Roadmap

1. Phase 1 (Current): Text-to-Speech implementation with basic configuration options.
2. Phase 2 (Future): Speech-to-Text and duplex conversation support.

Getting Started

Prerequisites

• Node.js 18.x
• Claude Desktop

Installation

• Clone the repository.
• Install dependencies.
• Start the server.

Integration Methods

1. Claude Desktop UI.
2. Command Line.
3. Using npx.

Technical Architecture

• MCP Protocol Implementation
• TTS Service with Google's API
• Session Management

API Documentation

REST Endpoints

• GET /api/v1/info: Server info
• GET /api/v1/health: Health check

WebSocket Messages

• tts_request: Text-to-speech requests

Troubleshooting

• Connection issues
• TTS Issues
• Integration Issues

Contributing

Contributions are welcome via Pull Requests.

License

MIT License.