mcp-garmin-connect

LLM-agnostic Garmin Connect MCP server for training, recovery, sleep, stress, VO2 Max, race predictions, and running summaries.

Use the same local Garmin tools from MCP clients such as Claude Desktop, Claude Code, Cursor, Continue, Hermes, or from provider-backed demo agents using DeepSeek, OpenAI, or Claude API.

Why This Exists

Most fitness advice from LLMs is generic. This server lets an LLM inspect your real Garmin Connect data first, then answer with context from your activities, recovery markers, HRV, sleep, stress, VO2 Max, and running history.

The core server is provider-neutral. DeepSeek, OpenAI, and Claude integrations are optional demo bridges for clients that do not speak MCP directly.

Related MCP server: claude-garmin

Features

• Local MCP server for Garmin Connect over stdio or Streamable HTTP.

• 14 Garmin tools and 4 prompt templates.

• Provider demo bridge for deepseek, openai, and claude.

• Windows-first CLI, with cross-platform Python project structure.

• Explicit opt-in before Garmin health/activity data is sent to any external LLM provider.

• Mocked test suite plus optional live Garmin smoke tests.

Languages And Stack

Quick Start

uv sync
Copy-Item .env.example .env
notepad .env

Set Garmin credentials:

GARMIN_EMAIL=you@example.com
GARMIN_PASSWORD=your-password

Authenticate once:

uv run garmin-mcp login
uv run garmin-mcp doctor --live

List available MCP tools:

uv run garmin-mcp tools

Run as a local MCP server:

uv run garmin-mcp serve --transport stdio

Run as an HTTP MCP server:

uv run garmin-mcp serve --transport http --host 127.0.0.1 --port 8765

HTTP clients can connect to:

http://127.0.0.1:8765/mcp

Provider Demo Agents

Provider API access may require a paid account or credits even if the provider's web app has a free tier.

DEEPSEEK_API_KEY=sk-...
OPENAI_API_KEY=sk-...
ANTHROPIC_API_KEY=sk-ant-...

Ask through a provider bridge:

uv run garmin-mcp ask "How is my recovery today?" --provider deepseek --allow-external-health-data
uv run garmin-mcp ask "Summarize my running for 90 days." --provider openai --allow-external-health-data
uv run garmin-mcp ask "What should I focus on this week?" --provider claude --allow-external-health-data

The --allow-external-health-data flag is required because Garmin tool results can be sent to the selected provider.

MCP Tools

MCP Prompts

• recovery_check

• weekly_training_review

• activity_analysis

• race_plan_context

Client Setup

See docs/clients for examples covering:

• Claude Desktop

• Claude Code

• Cursor

• Continue

• Hermes

• Generic stdio MCP clients

• Generic HTTP MCP clients

Minimal stdio config:

{
  "mcpServers": {
    "garmin": {
      "command": "uv",
      "args": [
        "--directory",
        "C:\\Users\\Jason\\Documents\\GitHub\\mcp-garmin",
        "run",
        "garmin-mcp",
        "serve",
        "--transport",
        "stdio"
      ]
    }
  }
}

Testing

uv run garmin-mcp doctor
uv run garmin-mcp doctor --live
uv run garmin-mcp tools
uv run ruff check .
uv run pytest

Optional live tests are gated and do not run in CI by default:

$env:GARMIN_LIVE_TEST = "1"
uv run pytest tests/live

Current local validation target:

ruff: All checks passed
pytest: all mocked tests pass, live test skipped unless enabled

Privacy And Safety

Your .env file and Garmin token cache stay local and are ignored by git.

MCP clients that run locally can call Garmin tools without sending provider API keys to this project. Provider demo commands such as garmin-mcp ask can send Garmin health/activity results to the selected LLM provider, so the CLI requires --allow-external-health-data.

Garmin Disclaimer

This project uses the community garminconnect Python package and Garmin Connect endpoints. Garmin Connect is not a public API for individual open-source projects, and authentication or endpoint behavior may change without notice.

Some metrics depend on Garmin device capabilities. For example, a device may expose VO2 Max and Training Effect but not native Garmin Training Load or Training Status.

Star History

After this repository is published, open the Star History page above and use its Show real-time chart on your README.md generator to create the live README embed. Star History currently requires a GitHub token for live charts because GitHub restricts star history data to repository owners and collaborators.

License

This project is licensed under the MIT License.

Publish

Codex does not push this repo. To publish it yourself:

git remote add origin https://github.com/Kavakoss/mcp-garmin-connect.git
git push -u origin main

Made by kavakoss.

Inspired by Jack-Abyss/claude-garmin, with a clean-room implementation focused on LLM-agnostic MCP usage.