Back to projects
Engineering

GroupMe MCP Server

MCP server wrapping the GroupMe v3 API — lets Claude and other AI assistants read messages, send to groups, and manage direct chats natively.

PythonMCPAI ToolsAPI Integration

Overview

An MCP (Model Context Protocol) server that wraps the GroupMe v3 REST API, making group messaging accessible to Claude and any other MCP-compatible AI assistant as a native tool.

Problem

GroupMe has no official AI integration. Asking Claude to read or send GroupMe messages requires manually copying message content in and out of the conversation — a friction point for any workflow that involves group coordination.

Approach

MCP Tool Surface

The server exposes GroupMe's core operations as MCP tools:

  • List groups — returns all group memberships with names and IDs
  • Read messages — fetches recent messages from a group
  • Send message — posts to a group
  • List direct chats — returns active DM threads
  • Send direct message — posts to a specific user

Authentication

A GroupMe access token is passed via environment variable at server startup. The token is obtained from the GroupMe developer portal in seconds — no OAuth flow required.

Installation

Configured in Claude's MCP settings as a local server. Once installed, Claude can read and send GroupMe messages directly within any conversation.

Tech Stack

  • Python with the mcp SDK
  • GroupMe v3 API via HTTP requests
  • uv for dependency management

Results

Claude can now participate in GroupMe group coordination natively — reading thread context, drafting replies, and sending messages without leaving the conversation.