Voice Call MCP Server

What is Voice Call MCP Server?

Voice Call MCP Server is a Model Context Protocol (MCP) server that enables Claude and other AI assistants to initiate and manage voice calls using Twilio and OpenAI’s GPT-4o Realtime model. It serves as a base for building AI-powered voice calling applications.

How to use Voice Call MCP Server?

Clone the repository, install dependencies (npm install && npm run build), and configure environment variables (Twilio credentials, OpenAI API key, ngrok authtoken). Add the server configuration to Claude Desktop’s claude_desktop_config.json, restart Claude, and you’ll see “Voice Call” under the hammer menu. Then invoke call tasks naturally, e.g., “Can you call +1-123-456-7890 and let them know I’ll be 15 minutes late?”

Key features of Voice Call MCP Server

• Make outbound phone calls via Twilio
• Process call audio in real-time with GPT-4o Realtime model
• Real-time language switching during calls
• Pre-built prompts for common scenarios (e.g., restaurant reservations)
• Automatic public URL tunneling with ngrok
• Secure handling of credentials

Use cases of Voice Call MCP Server

• Notify someone about a meeting delay via a simple phone call
• Make a restaurant reservation in a specified language
• Reschedule an appointment by speaking to a business

FAQ from Voice Call MCP Server

What are the requirements to run Voice Call MCP Server?

Node.js 22 or later, a Twilio account with API credentials, an OpenAI API key, and an ngrok authtoken. Using nvm to manage Node.js versions is recommended.

How do I configure Voice Call MCP Server for Claude Desktop?

Set environment variables for TWILIO_ACCOUNT_SID, TWILIO_AUTH_TOKEN, TWILIO_NUMBER, OPENAI_API_KEY, and NGROK_AUTHTOKEN in the claude_desktop_config.json file. Restart Claude Desktop to load the configuration.

What should I know about phone number format, limits, and costs?

Phone numbers must be in E.164 format (e.g., +11234567890). Be aware of Twilio and OpenAI account rate limits and pricing—call durations affect both services. The ngrok tunnel exposes your server publicly with a random URL and a random secret.

How is data handled and secured?

The server is open-source and designed for transparency and customizability. Credentials are stored securely in environment variables. No local database is included; real-time audio flows through Twilio and OpenAI. Call recording is optional via the RECORD_CALLS environment variable.