PhoneBooth

You can equip an AI agent with real-time phone capabilities by running the PhoneBooth MCP Server. This server exposes a local MCP interface that enables calling, monitoring, and transcribing phone calls through AI-driven workflows. It’s designed for easy integration with your existing AI agents and MCP clients, using a simple command-line startup and an API key for authentication.

How to use

You use the PhoneBooth MCP Server by running the provided MCP command on your machine and providing your API key. Then connect your MCP client or agent to the server to initiate calls, check status, and retrieve transcripts and summaries. The server exposes a set of tools that let you register, place calls, monitor progress, and download transcripts, all within your AI-enabled workflow.

How to install

Prerequisites: you need Node.js installed on your system. The server is distributed as an npm package you install globally.

Additional sections

Configuration and startup are straightforward. You generate or obtain an API key, then run the MCP server with that key supplied as an environment variable. You can also configure your MCP client to start the server with the appropriate environment variable so nothing else needs to be changed in your client code.

Example API key setup and startup sequences are provided below to help you get started quickly.

Available tools

register

Register and obtain an API key (one-time operation)

make_call

Place a call to a phone number using AI voice control

check_call

Retrieve call status, transcript, and summary

demo_call

Run a free demo call against simulated businesses

get_balance

Check remaining credits and plan tier

get_transcript

Download the full call transcript