ElevenLabs MCP Server

How to use

You will start the MCP server locally or in a remote environment, then connect an MCP client to perform operations such as creating and listing agents, registering tools, uploading knowledge base documents, and computing RAG indices. Use the client to manage resources and subscribe to real-time updates so you stay informed about changes to agents, tools, or knowledge bases. You can also enable Claude Desktop integration to work with Claude alongside your MCP setup.

Typical usage patterns include: creating an agent with a defined conversational configuration, adding tools the agent can call, uploading or scraping knowledge base documents for retrieval, and triggering a RAG index to improve answer quality. You can update or delete resources as needs change and monitor changes via real-time update subscriptions.

How to install

Prerequisites: you need Python and pip installed on your system. You should also have a working ElevenLabs API key for integration.

# Local development setup steps
# 1) Create and navigate into the project directory
# 2) Install project dependencies
pip install -r requirements.txt

# 3) Set up environment variables
cp .env.example .env
# Edit .env to include your ElevenLabs API key and other settings

# 4) Install the package in editable mode
pip install -e .

Production installation is supported by installing the MCP server package directly into your environment.

Configuration

Environment variables you should set in a .env file to configure the MCP server and its ElevenLabs integration include the following.

ELEVENLABS_API_KEY=your-elevenlabs-api-key-here
ELEVENLABS_BASE_URL=https://api.elevenlabs.io/v1
MCP_SERVER_NAME=elevenlabs-mcp-server
MCP_SERVER_VERSION=1.0.0
REQUEST_TIMEOUT=30
MAX_RETRIES=3
LOG_LEVEL=INFO

Claude Desktop Integration

To enable Claude Desktop support, add an MCP server entry with the Python runtime and the MCP module. Include your API key in the environment variables for Claude's configuration.

{
  "mcpServers": {
    "elevenlabs": {
      "command": "python",
      "args": ["-m", "elevenlabs_mcp.server"],
      "env": {
        "ELEVENLABS_API_KEY": "your-elevenlabs-api-key-here"
      }
    }
  }
}

Starting the server

Start the MCP server using the installed command or by invoking Python to run the MCP module.

elevenlabs-mcp-server

python -m elevenlabs_mcp.server

Additional notes

The MCP server provides resources for agents, tools, and knowledge bases. You can manage these resources via the MCP client using a consistent set of operations, such as creating, retrieving, listing, updating, and deleting resources. Real-time updates keep you informed of changes, and you can compute RAG indices to improve retrieval quality.

create_agent

Create a new conversational AI agent with a configuration that defines language, prompts, and initial messages.

get_agent

Retrieve the configuration for a specific agent by its identifier.

list_agents

List all agents with optional pagination or filtering.

update_agent

Update the configuration of an existing agent.

delete_agent

Remove an agent from the server.

create_tool

Create webhook or client-side tools that agents can invoke.

get_tool

Retrieve a tool configuration by its identifier.

list_tools

List tools with optional filtering.

update_tool

Update an existing tool configuration.

delete_tool

Delete a tool from the server.

create_knowledge_base_from_text

Create a knowledge base from plain text content.

create_knowledge_base_from_url

Create a knowledge base by scraping content from a URL.

get_knowledge_base_document

Retrieve details of a knowledge base document.

list_knowledge_base_documents

List all knowledge base documents.

update_knowledge_base_document

Update metadata for a knowledge base document.

delete_knowledge_base_document

Delete a knowledge base document.

compute_rag_index

Compute a Retrieval-Augmented Generation index for improved retrieval.

get_document_content

Fetch the full document content and its chunks.