health-mcp
Exposes personal Garmin wellness data through MCP tools for accessing summary, sleep, HRV, heart rate, stress, body battery, and historical data.
README
<div align="center"> <img src="icon.png" alt="health-mcp logo" width="140"> <h1>health-mcp</h1> <p>An MCP server that exposes personal Garmin wellness data.</p> </div>
Setup
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
export GARMIN_EMAIL="you@example.com"
export GARMIN_PASSWORD="your-password"
export GARMIN_TOKEN_STORE="$HOME/.config/health-mcp/garmin-tokens"
export GARMIN_CACHE_TTL_SECONDS="300"
python server.py
The same variables are listed in .env.example. Copy it to .env and replace
the example values. The server loads .env from the project directory.
GARMIN_TOKEN_STORE is optional and defaults to the path shown above. Login
tokens are reused so the account credentials are not submitted on every call.
Garmin responses are cached in memory for five minutes by default. Set
GARMIN_CACHE_TTL_SECONDS to change the duration, or set it to 0 to disable
caching. The cache is shared across tools and resets when the server restarts.
Failed Garmin requests are never cached.
If Garmin requests MFA on the first login, set GARMIN_MFA_CODE to the current
code and restart the server. Remove it after tokens have been created.
For Garmin China accounts, set GARMIN_IS_CN=true.
Running
For an MCP client that launches the server as a subprocess:
python server.py
This uses stdio. Do not type tool names into that terminal because standard
input carries MCP JSON-RPC messages.
To expose a local Streamable HTTP MCP endpoint:
python server.py --transport http
The endpoint is http://127.0.0.1:8000/mcp. It is an MCP endpoint rather than
a REST endpoint, so connect with an MCP client. Use --host and --port to
change the binding.
Project structure
server.py # Thin command-line entry point
gateway.py # Garmin Connect integration
mcp/
├── main.py # Server construction and transport startup
├── models.py # MCP response models
└── tools.py # MCP tool registration and handlers
tests/ # Unit tests
The local mcp/ directory intentionally has no __init__.py. FastMCP depends
on a third-party Python package also named mcp, so making this directory a
package would shadow that dependency.
Tools
get_summaryget_sleepget_hrvget_heart_rateget_stressget_body_batteryget_history
All tools accept ISO dates (YYYY-MM-DD) and default to the current local date.
The Body Battery tool accepts an optional date range.
get_history accepts an inclusive start_date, optional end_date, and an
optional list of metrics:
{
"start_date": "2026-06-01",
"end_date": "2026-06-07",
"metrics": ["sleep", "hrv", "heart_rate"]
}
History requests are limited to 31 days. Body Battery is fetched with Garmin's
range endpoint. Summary, sleep, HRV, heart rate, and stress use Garmin's daily
endpoints in date order. Successful dates are returned even when another date
fails; failures appear in the response's errors list. Cached daily responses
are reused across overlapping history requests.
This project uses Garmin Connect's unofficial API. Garmin may change its endpoints or response fields without notice.
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.