garmin-mcp
A local MCP server that exposes Garmin Connect activity, health, and training data as tools for Claude Desktop. It provides read-only access to metrics like activities, sleep, heart rate, and training readiness.
README
garmin-mcp
A local MCP server that exposes your Garmin Connect activity, health, and training data (activities, HRV, sleep, heart rate, stress, body battery, training readiness, HR/power zones, lactate threshold, FTP, VO2max, personal records, workouts, scheduled workouts, and training plans) as tools inside Claude Desktop.
It talks directly to the live Garmin Connect API via the
garminconnect
library. Data is read-only — nothing is ever written back to Garmin.
How auth works
Garmin login happens in two separate places on purpose:
scripts/setup_auth.py— an interactive script you run yourself in a terminal. It logs in with your email/password, prompts you for an MFA code if Garmin asks for one, and saves the resulting session token to~/.garmin_mcp/tokens.- The MCP server (
src/garmin_mcp/server.py), launched headlessly by Claude Desktop, only ever loads and silently refreshes that cached token. It never prompts for MFA — Claude Desktop gives it no terminal to prompt on, and blocking on stdin there would hang tool calls. If the cached session is missing or has expired, tool calls fail with an error telling you to re-runsetup_auth.py.
Your Garmin password is stored in the macOS keychain (item/service
garmin-mcp, viewable in Keychain Access), never written to a file. Only
your Garmin login email (not secret) is saved, in ~/.garmin_mcp/config.json.
Setup
-
Install dependencies:
cd garmin_mcp uv sync -
Run the one-time interactive login:
uv run python scripts/setup_auth.pyEnter your Garmin Connect email and password (the password is then stored in the keychain for next time). Enter the MFA code when prompted. On success you'll see a confirmation and the token cache path.
-
Register the server with Claude Desktop by editing
~/Library/Application Support/Claude/claude_desktop_config.jsonand adding (alongside any othermcpServersentries):{ "mcpServers": { "garmin": { "command": "uv", "args": ["run", "--project", "/Users/felix.guiony/Code/Perso/garmin_mcp", "garmin-mcp"] } } } -
Fully quit Claude Desktop (Cmd+Q, not just close the window) and reopen it. The tools/hammer icon should list the
garmintools.
Tools
| Tool | Description |
|---|---|
list_activities |
Recent activities, most recent first (paginated, optional type filter) |
get_activity |
Full detail for one activity: summary + per-metric time series |
get_activity_splits |
Lap/split data for one activity |
get_heart_rate |
Daily resting HR + intraday HR timeline |
get_daily_stats |
Steps, calories, resting HR, distance, floors, intensity minutes |
get_sleep |
Sleep stages, duration, sleep score |
get_hrv |
Heart rate variability |
get_body_battery |
Body Battery energy reserve over a date range |
get_stress |
All-day stress level |
get_training_readiness |
Training Readiness score and contributing factors |
get_profile |
Identity, unit prefs, personal settings (max HR, resting HR, weight, height, VO2max), HR zones and power zones per sport |
get_lactate_threshold |
Running lactate threshold: heart rate, power, and speed |
get_cycling_ftp |
Latest cycling Functional Threshold Power |
get_ftp_history |
Historic FTP for a sport over a date range |
get_max_metrics |
Max-metric data (e.g. VO2max) over a date range |
get_resting_heart_rate |
Daily resting heart rate trend over a date range |
get_fitness_age |
Garmin Fitness Age for a given date |
get_personal_records |
Personal records |
list_workouts |
Saved workout templates, most recent first (paginated) |
get_workout |
Full structure of one workout: segments, targets, intervals |
get_scheduled_workouts |
Calendar of workouts scheduled for a given month |
get_scheduled_workout |
Detail for one scheduled workout instance |
list_training_plans |
Training plans |
get_training_plan |
Details for a specific training plan |
get_adaptive_training_plan |
Details for a specific adaptive training plan |
All date parameters take YYYY-MM-DD and default to today.
Troubleshooting
A tool call returns a "Garmin session error" / "run setup_auth.py" message. Your cached refresh token expired or was revoked (this happens occasionally, e.g. after a password change or long period of inactivity). Re-run:
uv run python scripts/setup_auth.py
Rate limited. Garmin Connect will occasionally rate-limit rapid requests; the underlying library retries transient failures automatically, but if you see a rate-limit error, wait a bit and try again.
Verifying the server directly, outside Claude Desktop:
uv run mcp dev src/garmin_mcp/server.py
opens the MCP Inspector so you can call each tool by hand and check its output/schema.
Project layout
garmin_mcp/
pyproject.toml
src/garmin_mcp/
auth.py # keychain + token-cache helpers
server.py # FastMCP server and tool definitions
scripts/
setup_auth.py # interactive one-time/occasional login
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.