elevenlabs/elevenlabs-mcp

What it solves

This project provides a Model Context Protocol (MCP) server that allows AI clients (such as Claude Desktop, Cursor, and Windsurf) to directly interact with ElevenLabs' audio APIs. It bridges the gap between LLMs and professional-grade text-to-speech, voice cloning, and audio processing tools, enabling AI agents to generate speech and manipulate audio files.

How it works

The server acts as a middleware that exposes ElevenLabs' API capabilities as tools to an MCP-compatible client. Once configured with an API key, the AI client can call specific functions to perform tasks like generating speech, transcribing audio, or cloning voices. It supports three different output modes for handling generated audio files: saving them to the local disk, returning them as base64-encoded MCP resources, or doing both.

Who it’s for

Developers and AI power users who want to integrate high-quality voice synthesis and audio manipulation into their AI agent workflows or IDEs.

Highlights

• Broad Client Support: Compatible with Claude Desktop, Cursor, Windsurf, and OpenAI Agents.
• Diverse Audio Capabilities: Supports text-to-speech, voice cloning, speech-to-text transcription, and soundscape generation.
• Flexible File Handling: Configurable output modes (files, resources, or both) to accommodate local or cloud-based environments.
• Security Boundaries: Implements a base path configuration to restrict file operations to a specific directory.