garmin-mcp

Enables Claude and other MCP-compatible AI models to establish bidirectional communication with Garmin wearables and fitness devices through the Model Context Protocol. Implements MCP server architecture that translates Garmin device APIs into standardized tool definitions, allowing language models to query real-time health metrics, activity data, and device status without direct API integration overhead.

Provides structured access to current and historical activity data from paired Garmin devices including steps, heart rate, sleep metrics, stress levels, and workout summaries. Implements query patterns that map natural language requests to Garmin Connect API endpoints, returning parsed JSON responses with typed fields for metrics like calories burned, distance, elevation gain, and biometric data.

Enables querying and managing multiple paired Garmin devices through a single MCP interface, providing device discovery, status monitoring, and device-specific capability detection. Implements device registry patterns that cache device metadata and capabilities, allowing AI models to understand which metrics are available per device and route queries appropriately.

Leverages MCP's integration with Claude and other language models to provide natural language interpretation of Garmin metrics, translating raw numbers into actionable insights. Works by exposing structured fitness data through MCP tools, allowing the AI model's reasoning capabilities to analyze trends, identify patterns, and generate personalized health recommendations based on the retrieved data.

Integrates Garmin fitness data as contextual information within MCP's resource system, allowing AI agents to automatically consider user health status when making decisions or recommendations. Implements context injection patterns where relevant Garmin metrics are retrieved and included in the model context window, enabling agents to factor in current activity levels, sleep quality, stress levels, and recovery status into their reasoning.