Taxuspt/garmin_mcp

What it solves

This project provides a bridge between Garmin Connect and AI assistants, allowing you to query your fitness data, manage workouts, and analyze performance metrics directly through an LLM interface like Claude Desktop.

How it works

It implements the Model Context Protocol (MCP) to expose over 110 Garmin Connect API endpoints as tools. By authenticating once via a secure CLI flow, the server stores OAuth tokens locally, enabling the LLM to securely access your health and activity data without requiring credentials to be stored in your configuration files. It supports both reading data (like heart rate, sleep, and training status) and writing data (like editing activity details or scheduling workouts).

Who it’s for

This is for users of Garmin fitness devices who want to use AI to gain deeper insights into their training, automate the creation of complex workout schedules, or perform advanced analysis on their activity data using natural language.

Highlights

• Extensive Tooling: Provides 110+ tools covering activities, health metrics, training status, gear management, and nutrition.
• Advanced Workout Builders: Allows you to create and schedule complex workouts (including intervals and strength training) without manually writing JSON.
• Secure Authentication: Uses a dedicated CLI tool to handle OAuth tokens, keeping sensitive credentials out of your MCP configuration.
• Customizable Access: Supports tool filtering (allowlists/denylists) to limit which data or actions the AI can access, and supports multiple Garmin accounts simultaneously.