Zonos TTS MCP for Linux

Zonos TTS MCP for Linux

A Model Context Protocol integration for Zonos TTS, allowing Claude to generate speech directly on Linux systems.

Overview

This project enables text-to-speech capabilities for Claude AI using the Zonos TTS system. The integration uses the Model Context Protocol (MCP) to allow Claude to generate natural-sounding speech in multiple languages and emotions.

Key Features

• Text-to-speech through Claude
• Multiple emotions support (neutral, happy, sad, angry)
• Multi-language support
• Proper audio playback through PulseAudio/PipeWire
• Compatible with Linux environments

Requirements

• Node.js
• PulseAudio or PipeWire with PulseAudio compatibility layer
• Running instance of Zonos API (PhialsBasement/zonos-api)
• Working audio output device
• Claude Desktop app

Setup

Installing via Smithery

To install the original UNMODIFIED version of Zonos TTS Integration for Claude Desktop automatically via Smithery:

npx -y @smithery/cli install @PhialsBasement/zonos-tts-mcp --client claude

Manual installation for Linux with GPU optimizations

1. Make sure you have Zonos API running with the API implementation (PhialsBasement/zonos-api)

2. Clone this repository:

git clone https://github.com/angrysky56/Zonos-TTS-MCP-Linux.git
cd Zonos-TTS-MCP-Linux

3. Install dependencies:

npm install @modelcontextprotocol/sdk axios zod

4. Build the MCP server:

npm run build

5. Add to Claude's config file: Edit your Claude config file (search for files containing "mcpServers" in your home directory) and add this to the mcpServers section:

"zonos-tts-mcp": {
  "command": "node",
  "args": [
    "/path/to/your/Zonos-TTS-MCP-Linux/dist/server.js"
  ]
}

Replace /path/to/your/Zonos-TTS-MCP-Linux with the actual path where you installed this project.

Troubleshooting

Error Messages in Console

If you encounter error messages like:

Unexpected token 'P', 'Playing au'... is not valid JSON

This project includes fixes for these common issues:

1. Console output has been redirected to a log file (/tmp/zonos-tts-mcp.log) instead of standard output to prevent JSON parsing errors.
2. The TTS parameters have been optimized for better performance.

Performance Optimization

This fork includes performance optimizations to reduce GPU load:

• Using more efficient sampling parameters (top_p and min_p)
• Improved log handling to prevent interference with MCP communication

Using with Claude

Once configured, Claude can use the speak_response tool:

speak_response(
    text="Your text here",
    language="en-us",  # optional, defaults to en-us
    emotion="happy"    # optional: "neutral", "happy", "sad", "angry"
)

Credit

This project is a modified version of PhialsBasement/zonos-tts-mcp, adapted for improved Linux compatibility and performance.

The Zonos TTS system is developed by Zyphra and is available through their Zonos API.

License

See the original project for license information.

Notes

• Make sure both the Zonos API server and this MCP server are running
• Audio playback requires proper PulseAudio/PipeWire configuration
• If you experience high GPU usage, you may want to adjust the TTS parameters in the server code