home / mcp / retell ai mcp server
Retell Simp MCP server
Configuration
View docs{
"mcpServers": {
"itsanamune-retellsimp": {
"command": "node",
"args": [
"/path/to/retell-ai-mcp/dist/index.js"
],
"env": {
"RETELL_API_KEY": "your-api-key-here"
}
}
}
}You can run and interact with the Retell AI API through a dedicated MCP server that exposes a suite of tools for building, testing, deploying, and monitoring AI phone agents. With this server you can manage calls, chats, phone numbers, agents, knowledge bases, and more, all from MCP-compatible clients like Claude. This guide shows you how to install, configure, and use the server end to end.
You will connect a Claude (or any MCP-compatible) client to the Retell AI MCP server to perform common operations. Start by ensuring the server is running with your API key, then call tools to create and manage calls, chats, numbers, agents, and knowledge bases. You can orchestrate conversations, schedule batch tasks, and publish agent versions from your MCP client by selecting the appropriate tool and supplying the required parameters. Use natural language prompts or tool-specific actions in your client to list resources, retrieve details, or trigger actions such as creating a phone call, starting a chat, or publishing an agent.
Common usage patterns include: finding available voices, creating a voice agent, setting up knowledge bases, and designing LLM configurations and conversation flows. Your MCP client will present the tool set, and you will provide inputs such as IDs, names, prompts, and configuration options to complete each operation.
Prerequisites you need before starting: Node.js and npm (or an environment that supports running MCP stdio servers). Ensure your environment can run JavaScript/TypeScript builds.
Step 1: Install dependencies and build the MCP server.
Step 2: Start the MCP server as described in the configuration steps below.
Environment variable you must set: RETELL_API_KEY with your Retell AI API key. This key authenticates requests to Retell AI and enables the MCP server to access endpoints. You can obtain the key from your Retell AI Dashboard.
Two common ways to wire up the MCP server with Claude Code: 1) CLI method and 2) Project configuration file. The CLI method is recommended for quick setup, while the project file helps share configuration across teams.
CLI method: add the MCP server using a stdio transport and provide the API key in the environment.
claude mcp add --transport stdio retell-ai -e RETELL_API_KEY=your-api-key-here -- node /path/to/retell-ai-mcp/dist/index.js
```, and you can also use environment variable expansion.Environment-variable expansion variant: use the existing RETELL_API_KEY in your shell.
claude mcp add --transport stdio retell-ai -e RETELL_API_KEY=${RETELL_API_KEY} -- node /path/to/retell-ai-mcp/dist/index.js
```, this keeps the key out of the command line and reads it from your environment.Project configuration file: create a .mcp.json at your project root to share with teammates.
{
"mcpServers": {
"retell-ai": {
"command": "node",
"args": ["/path/to/retell-ai-mcp/dist/index.js"],
"env": {
"RETELL_API_KEY": "${RETELL_API_KEY}"
}
}
}
}
```, this expands the environment variable at runtime.Add the MCP server configuration to Claude Desktop so you can access Retell AI tools from your desktop client. Use the file path appropriate to your OS: macOS or Windows.
{
"mcpServers": {
"retell-ai": {
"command": "node",
"args": ["/path/to/retell-ai-mcp/dist/index.js"],
"env": {
"RETELL_API_KEY": "your-api-key-here"
}
}
}
}
```, place this in the Claude Desktop config at the following paths: macOS: ~/Library/Application Support/Claude/claude_desktop_config.json Windows: %APPDATA%\Claude\claude_desktop_config.jsonYou can, for instance, list available voices, create a voice agent, and schedule a batch of outbound calls using the Retell AI MCP server through your MCP client.
Create a new outbound phone call
Create a new web call session
Retrieve details of a specific call
List and filter calls
Update call metadata
Delete a call
Create a chat session
Start an SMS conversation
Get chat details
Send a message in a chat and receive a response
List all chats
End a chat session
Purchase a phone number
Get phone number details
List all phone numbers
Update phone number settings
Delete a phone number
Import a phone number via SIP
Create a voice agent
Get agent details
List all agents
Update agent config
Delete an agent
Publish agent version
Get version history of an agent
Create a chat agent
Get chat agent details
List chat agents
Update chat agent
Delete chat agent
Create LLM configuration
Get LLM details
List all LLMs
Update LLM config
Delete LLM
Create a conversation flow
Get flow details
List all conversation flows
Update a conversation flow
Delete a conversation flow
Create knowledge base
Get knowledge base details
List all knowledge bases
Delete knowledge base
Add sources to a knowledge base
Remove a knowledge base source
Get voice details
List available voices
Schedule bulk calls
Run agent tests
Check concurrent call limits
Register inbound call