Retell AI MCP Server

A Model Context Protocol (MCP) server that provides tools for interacting with the Retell AI API. Build, test, deploy, and monitor AI phone agents directly through Claude or other MCP-compatible clients.

Features

This MCP server exposes 45+ tools covering all major Retell AI API endpoints:

Call Management

• Create outbound phone calls and web calls
• Retrieve call details including transcripts and recordings
• List and filter calls with pagination
• Update call metadata and data storage settings
• Delete calls and associated data

Chat Management

• Create chat sessions and SMS conversations
• Send messages and receive responses
• List and manage chat sessions

Phone Number Management

• Purchase/register new phone numbers
• Import existing numbers via SIP
• Configure inbound/outbound agents per number
• List and manage phone numbers

Agent Management

• Create and configure voice agents
• Create and configure chat agents
• Publish agent versions
• Manage agent settings (voice, language, behavior)

LLM & Conversation Flow

• Create and manage Retell LLM configurations
• Design conversation flows with node-based structure
• Configure custom prompts and function calling

Knowledge Base

• Create knowledge bases for agent context
• Add sources (URLs, text, documents)
• Manage knowledge base sources

Voice & Batch Operations

• List available voices
• Schedule batch outbound calls
• Run automated agent tests

Account

• Check concurrent call limits

Installation

npm install
npm run build

Configuration

Environment Variable

Set your Retell AI API key as an environment variable:

export RETELL_API_KEY="your-api-key-here"

Get your API key from the Retell AI Dashboard.

Claude Code Configuration

Option 1: Using the CLI (Recommended)

Add the MCP server using the claude mcp add command:

claude mcp add --transport stdio retell-ai -e RETELL_API_KEY=your-api-key-here -- node /path/to/retell-ai-mcp/dist/index.js

You can also use environment variable expansion:

claude mcp add --transport stdio retell-ai -e RETELL_API_KEY=\${RETELL_API_KEY} -- node /path/to/retell-ai-mcp/dist/index.js

Option 2: Project Configuration File

Create a .mcp.json file in your project root to share with your team:

{
  "mcpServers": {
    "retell-ai": {
      "command": "node",
      "args": ["/path/to/retell-ai-mcp/dist/index.js"],
      "env": {
        "RETELL_API_KEY": "${RETELL_API_KEY}"
      }
    }
  }
}

The ${RETELL_API_KEY} syntax will expand the environment variable at runtime.

Managing the Server

# List configured MCP servers
claude mcp list

# Get details about the retell-ai server
claude mcp get retell-ai

# Remove the server
claude mcp remove retell-ai

Claude Desktop Configuration

Add this to your Claude Desktop config file:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "retell-ai": {
      "command": "node",
      "args": ["/path/to/retell-ai-mcp/dist/index.js"],
      "env": {
        "RETELL_API_KEY": "your-api-key-here"
      }
    }
  }
}

Available Tools

Call Tools

Chat Tools

Phone Number Tools

Voice Agent Tools

Chat Agent Tools

LLM Tools

Conversation Flow Tools

Knowledge Base Tools

Voice Tools

Batch Tools

Account Tools

Example Usage

Once configured, you can use Claude to interact with Retell AI:

List available voices:

"Show me all available voices for Retell AI"

Create an agent:

"Create a new voice agent named 'Support Bot' using the voice '11labs-Adrian' with a friendly greeting"

Make a phone call:

"Call +1234567890 from my registered number +0987654321 using the Support Bot agent"

Check call history:

"Show me all calls from the last 24 hours"

Development

# Install dependencies
npm install

# Build TypeScript
npm run build

# Watch mode for development
npm run dev

License

MIT

Links

• Retell AI Documentation
• Retell AI Dashboard
• Claude Code MCP Documentation
• MCP Protocol