
Fitbit MCP Server
A Model Context Protocol server that bridges LLMs (like Claude) with the Fitbit API, allowing the LLM to request and retrieve user health and fitness data through defined tools.
Tools
get_weight
Get the raw JSON response for weight entries from Fitbit for a specified period ending today. Requires a 'period' parameter such as '1d', '7d', '30d', '3m', '6m', '1y'
get_sleep_by_date_range
Get the raw JSON response for sleep logs from Fitbit for a specific date range. Requires 'startDate' and 'endDate' parameters in 'YYYY-MM-DD' format. Note: The API enforces a maximum range of 100 days.
get_profile
Get the raw JSON response for the user's Fitbit profile.
README
Fitbit MCP Server
This project implements a Model Context Protocol (MCP) server that acts as a bridge between an LLM (like Claude) and the Fitbit API.
It allows the LLM to request and retrieve health and fitness data from a user's Fitbit account via defined tools.
Features
- Fitbit API Integration: Connects securely to the Fitbit API using OAuth 2.0.
- MCP Compliance: Exposes Fitbit data through tools compliant with the Model Context Protocol.
- Extensible: Designed to easily add support for more Fitbit API endpoints.
Supported Endpoints
- Weight:
get_weight
: Retrieves raw weight data for a specified period (1d
,7d
,30d
,3m
,6m
,1y
) ending today. Requires aperiod
parameter.
- Sleep:
get_sleep_by_date_range
: Retrieves raw sleep log data for a specific date range (max 100 days). RequiresstartDate
andendDate
parameters inYYYY-MM-DD
format.
- Profile:
get_profile
: Retrieves the user's Fitbit profile information, including personal details such as name, age, gender, height, weight, and account information.
Planned Endpoints
- Steps
- Heart Rate
- Activity
Setup
-
Clone the Repository:
git clone https://github.com/TheDigitalNinja/mcp-fitbit cd MCP-Server
-
Install Dependencies:
npm install
-
Create Environment File:
- Create a file named
.env
in the project root (mcp-fitbit/
). - Add your Fitbit application credentials:
FITBIT_CLIENT_ID=YOUR_FITBIT_CLIENT_ID FITBIT_CLIENT_SECRET=YOUR_FITBIT_CLIENT_SECRET
- You can obtain these by registering an application at dev.fitbit.com. Ensure the OAuth 2.0 Application Type is set to
Personal
and the Callback URL ishttp://localhost:3000/callback
.
- Create a file named
-
Build the Project:
npm run build
This compiles the TypeScript source files (
src/
) into JavaScript files (build/
).
Integrating with Claude for Desktop
To use this Fitbit MCP server with Claude for Desktop:
-
Locate the Claude Configuration File:
- On Windows, open the file explorer and navigate to
%AppData%\Claude\
. - Open the
claude_desktop_config.json
file in a text editor.
- On Windows, open the file explorer and navigate to
-
Add the Server Configuration:
- Find the
mcpServers
key in the JSON file (if it doesn't exist, you might need to add it). - Add an entry for the Fitbit server. Replace
C:\PATH\TO\PARENT\FOLDER\mcp-fitbit
with the absolute path to the directory where you cloned this repository.
{ "mcpServers": { "fitbit": { "command": "node", "args": [ "C:\\PATH\\TO\\PARENT\\FOLDER\\mcp-fitbit\\build\\index.js" ] } // Add other servers here if you have them } // ... other Claude settings ... }
- Important: Ensure you use double backslashes (
\\
) for paths on Windows in the JSON file.
- Find the
-
Save and Restart:
- Save the
claude_desktop_config.json
file. - Restart the Claude for Desktop application.
- Save the
-
First Run & Authorization:
- The first time Claude connects to the server (e.g., when you ask it a question that requires a Fitbit tool), the server will initiate the Fitbit OAuth 2.0 authorization flow.
- A console window for the server might appear briefly.
- The server will attempt to open
http://localhost:3000/auth
in your default web browser. - If the browser doesn't open automatically, navigate to that URL manually.
- Log in to your Fitbit account and grant the application permission for the requested scopes (Weight and Sleep).
- You will be redirected to
http://localhost:3000/callback
. - A success message will appear in the browser tab, which can then be closed.
- The server is now authenticated and Claude can use the Fitbit tools.
Available Tools
Once the server is running and authorized, the following tools will be available to the connected LLM:
get_weight
: Fetches raw weight time series data as a JSON string for a specified period ending today.- Parameter:
period
(string, required) - Specifies the duration. Must be one of:"1d"
,"7d"
,"30d"
,"3m"
,"6m"
,"1y"
. - Example Usage (Conceptual):
get_weight(period="7d")
- Parameter:
get_sleep_by_date_range
: Fetches raw sleep log data as a JSON string for a specified date range (maximum 100 days).- Parameters:
startDate
(string, required) - Specifies the start date inYYYY-MM-DD
format.endDate
(string, required) - Specifies the end date inYYYY-MM-DD
format.
- Example Usage (Conceptual):
get_sleep_by_date_range(startDate="2025-04-01", endDate="2025-04-30")
- Parameters:
get_profile
: Fetches the user's Fitbit profile information as a JSON string.- Parameters: None required.
- Example Usage (Conceptual):
get_profile()
Development
- Source code is located in the
src/
directory. - Run
npm run build
after making changes to the TypeScript files. - Restart Claude for Desktop to ensure it picks up the latest build if the server was already running via Claude.
推荐服务器

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。
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。
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。
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。

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。
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。
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。

e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
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。