ElevenLabs official MCP server, supporting AI audio processing functions such as text-to-speech, voice cloning, and audio transcription.

ElevenLabs MCP Server Detailed Introduction

Project Overview

The ElevenLabs MCP server is the official Model Context Protocol (MCP) server from ElevenLabs, supporting interaction with powerful text-to-speech and audio processing APIs. This server allows MCP clients such as Claude Desktop, Cursor, Windsurf, and OpenAI Agents to generate speech, clone voices, transcribe audio, and more.

Core Features

Key Features

• Text-to-Speech (TTS): Converts text into natural-sounding speech.
• Voice Cloning: Replicates and simulates specific voice characteristics.
• Audio Transcription: Converts audio files into text.
• Voice Design: Creates and customizes unique voice configurations.
• Audio Isolation: Separates specific sounds from complex audio.
• Sound Effect Generation: Creates various sound effects and soundscapes.

Supported Clients

• Claude Desktop
• Cursor
• Windsurf
• OpenAI Agents
• Other MCP-compatible clients

Installation and Configuration

Prerequisites

1. Obtain an API key from ElevenLabs (the free version provides 10k credits per month).
2. Install uv (Python package manager):

   curl -LsSf https://astral.sh/uv/install.sh | sh
   

Claude Desktop Configuration

In Claude > Settings > Developer > Edit Config > claude_desktop_config.json, add the following configuration:

{
  "mcpServers": {
    "ElevenLabs": {
      "command": "uvx",
      "args": ["elevenlabs-mcp"],
      "env": {
        "ELEVENLABS_API_KEY": "<insert-your-api-key-here>"
      }
    }
  }
}

Windows Users Note: You need to enable "Developer Mode" in Claude Desktop. Click "Help" in the top-left hamburger menu and select "Enable Developer Mode".

Other Client Configuration

For clients like Cursor and Windsurf:

pip install elevenlabs-mcp
python -m elevenlabs_mcp --api-key={{PUT_YOUR_API_KEY_HERE}} --print

Then paste the output configuration into the corresponding MCP client's configuration directory.

Usage Examples

After installation, you can try the following:

• "Create an AI agent that speaks like a film noir detective and can answer questions about classic movies."
• "Generate three voice variations for a wise ancient dragon character, and then I'll choose my favorite to add to the voice library."
• "Transform this recording of my voice to sound like a medieval knight."
• "Create a soundscape of a thunderstorm in a dense forest, including animal reactions to the weather."
• "Transcribe this speech to text, identify the different speakers, and then re-convert it back to speech using a unique voice for each person."

Development and Contribution

Running from Source

1. Clone the repository:

   git clone https://github.com/elevenlabs/elevenlabs-mcp
   cd elevenlabs-mcp
   

2. Create a virtual environment and install dependencies:

   uv venv
   source .venv/bin/activate
   uv pip install -e ".[dev]"
   

3. Configure environment variables:

   cp .env.example .env
   # Edit the .env file and add your API key
   

4. Run tests:

   ./scripts/test.sh
   # Or run with options
   ./scripts/test.sh --verbose --fail-fast
   

Debugging and Testing

• Install the server in Claude Desktop:

  mcp install elevenlabs_mcp/server.py
  

• Use MCP Inspector for local debugging:

  mcp dev elevenlabs_mcp/server.py
  

Log Location

Log file location for Claude Desktop runtime:

• Windows: %APPDATA%\Claude\logs\mcp-server-elevenlabs.log
• macOS: ~/Library/Logs/Claude/mcp-server-elevenlabs.log

Technical Architecture

The ElevenLabs MCP server is an abstraction layer for the ElevenLabs API, providing large language models with contextual access to the full ElevenLabs AI audio platform. It acts as a developer-friendly local interface, forwarding requests to ElevenLabs' cloud API.

Important Notes

1. Timeout Handling: Some ElevenLabs API operations (such as voice design and audio isolation) may take a long time to complete. Timeout errors may occur in MCP inspector development mode, but the task will still complete. This issue does not occur when using clients like Claude.

2. Path Configuration: You can specify the base path for the MCP server to find and output relative path files by adding the ELEVENLABS_MCP_BASE_PATH environment variable.

3. Error Troubleshooting: If you encounter an "MCP ElevenLabs: spawn uvx ENOENT" error, use the which uvx command to confirm the absolute path, and then use the full path in the configuration (e.g., "command": "/usr/local/bin/uvx").

Project Value

This project provides AI application developers with:

• Standardized audio processing interface
• Seamless speech synthesis integration
• Powerful audio analysis capabilities
• Cross-platform compatibility
• Enterprise-grade audio quality

Through the MCP protocol, developers can easily integrate ElevenLabs' high-quality audio processing capabilities into various AI applications, greatly simplifying the development process for audio-related functions.