This MCP server exposes ElevenLabs text-to-speech and audio processing capabilities to MCP clients. It allows tools such as voice cloning, speech generation, and audio transcription to be driven through the MCP interface, enabling integration with clients like Claude Desktop, Cursor, Windsurf, and OpenAI Agents. The server is configured to work with the ElevenLabs MCP package, which exposes an API surface for generating speech, manipulating voice characteristics, and handling audio processing tasks in an MCP-friendly way. To use it, provide your ElevenLabs API key in the environment configuration so the MCP client can authenticate with the ElevenLabs service, and point your MCP client to the ElevenLabs MCP configuration generated by the server.

Once configured, you can use the MCP client to issue prompts or requests that trigger the underlying ElevenLabs APIs—for example generating speech from text, cloning a voice, or transcribing audio. The client tools described in the ElevenLabs MCP docs translate into MCP commands that the server handles, returning results such as audio data, file references, or transcription text, depending on the operation and output mode chosen.

Prerequisites:

• Python (for uv) and the uv package manager installed on your system
• Internet access to install dependencies from PyPI
• An ElevenLabs API key for authenticating requests

Step-by-step installation:

1. Install uv (Python package manager) if you don’t have it:
   • On macOS/Linux: curl -LsSf https://astral.sh/uv/install.sh | sh
   • On Windows, use the uv installation method appropriate for your environment.
2. Install the ElevenLabs MCP package in editable mode (from PyPI): uv pip install -e ".[dev]" (If you’re cloning the repo locally, run this inside the repository directory after cloning.)
3. Copy the example environment configuration and add your API key: cp .env.example .env

   Edit .env and insert your ElevenLabs API key

4. Run tests to verify the setup (optional but recommended): ./scripts/test.sh

Configuration steps:

• Create or edit your Claude Desktop (or other MCP client) configuration to point at the ElevenLabs MCP server using the provided mcpServers entry (see mcp_config above).
• Ensure ELEVENLABS_API_KEY is set in your environment or in the client configuration as required by the MCP client you are using.

Notes and tips:

• Ensure you have a valid ELEVENLABS_API_KEY. The MCP server will rely on this key to access ElevenLabs services.
• You can customize output handling using environment variables like ELEVENLABS_MCP_OUTPUT_MODE to control how generated files are returned (files, resources, or both).
• If you encounter spawn or path issues with uvx, verify the absolute path to the uvx executable and set it explicitly in your MCP configuration.
• Data residency can be controlled with ELEVENLABS_API_RESIDENCY (default us); this is enterprise-only and should be set according to organizational requirements.
• For troubleshooting, consult Claude Desktop logs on Windows/macOS as documented in the project README, and verify that your API key has the appropriate permissions for the operations you’re attempting.