Garmin MCP Server

Garmin MCP Server

A local, read-only MCP server that allows Claude Desktop to access Garmin Connect data such as activities and recovery metrics, enabling AI-assisted running plan creation and adjustment.

Category
访问服务器

README

Garmin MCP server

A local, read-only MCP server that lets Claude Desktop retrieve data from an approved Garmin Connect Developer Program project. Claude can use the returned activities and recovery metrics to create or adjust a running plan.

What it exposes

  • garmin_connection_status — authorized user and granted permissions
  • garmin_get_data — one Garmin summary category for a recent upload window
  • garmin_get_training_context — activities, daily health, sleep, HRV, stress, and fitness metrics in one call

The server does not generate a plan itself. It gives Claude the source data; Claude performs the analysis and plan customization.

Prerequisites

  • Node.js 20 or newer
  • An approved Garmin Connect Developer Program application
  • OAuth 2.0 + PKCE enabled for that application
  • Ping/Pull access and the summary feeds you want enabled in Garmin's portal

Garmin's Health API is a partner API, not a public personal-account API. A normal Garmin username and password are not API credentials.

Install and authorize

npm install
cp .env.example .env

Put the client ID and client secret from Garmin in .env. Register this exact redirect URI in Garmin's developer portal:

http://127.0.0.1:8787/callback

Then authorize your Garmin account and build the server:

npm run auth
npm run build

Authorization writes .garmin-tokens.json with owner-only file permissions. Both that file and .env are excluded from Git.

Garmin may assign environment-specific URLs to an evaluation project. If its portal documentation gives you different authorization, token, or API URLs, set the corresponding overrides in .env.

Add it to Claude Desktop

Open Claude Desktop → Settings → Developer → Edit Config, then add:

{
  "mcpServers": {
    "garmin": {
      "command": "node",
      "args": [
        "/Users/abhilashaarohi/Projects/garmin-mcp-server/dist/index.js"
      ],
      "cwd": "/Users/abhilashaarohi/Projects/garmin-mcp-server"
    }
  }
}

If your config already contains other servers, add only the garmin entry inside its existing mcpServers object. Fully restart Claude Desktop, then ask:

Check my Garmin connection and summarize my recent running and recovery context. Before creating a plan, ask for my race distance, race date, weekly availability, current weekly mileage, injury history, and preferred long-run day. Use conservative progression and explain any recovery concerns.

Test without Claude

npm run inspect

The MCP Inspector lets you call each tool and view its raw result.

Garmin data limitations

Garmin's Health API query windows are based on upload time, not activity date. Each API request can cover at most 24 hours, so this server splits larger requests automatically.

Garmin retains uploaded summaries for only seven days. This local version can retrieve recent uploads but cannot provide reliable long-term history unless it is run regularly. A production version should:

  1. expose a public HTTPS webhook for Garmin Ping or Push notifications;
  2. fetch each notification immediately;
  3. upsert updates into a database by Garmin summary ID and start time; and
  4. have MCP tools query that database.

Garmin's developer portal can initiate backfills for older history. Backfill results are asynchronous and must also be captured through Ping or Push.

Privacy and safety

  • Keep .env and .garmin-tokens.json private.
  • The tools are read-only and do not upload workouts to Garmin.
  • Treat an AI-generated running plan as general training guidance, not medical advice. Pain, illness, abnormal cardiovascular symptoms, or injury requires appropriate professional advice.

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选