Get the FREE Ultimate OpenClaw Setup Guide →

mcp-fitbit

Give your AI assistant access to your Fitbit data for personalized health insights, trend analysis, and automated tracking. Works with Claude Desktop and other MCP-compatible AI tools.

Installation
Run this command in your terminal to add the MCP server to Claude Code.
Run in terminal:
Command
claude mcp add --transport stdio thedigitalninja-mcp-fitbit mcp-fitbit \
  --env FITBIT_CLIENT_ID="your_client_id_here" \
  --env FITBIT_CLIENT_SECRET="your_client_secret_here"

How to use

The Fitbit MCP Connector (mcp-fitbit) exposes a suite of tools that let your AI assistants query your Fitbit health data through MCP. It supports retrieving weight trends, sleep data, exercise logs, heart rate, nutrition, and basic profile information. Tools include get_weight for weight history, get_sleep_by_date_range for sleep logs, get_exercises for activity logs, get_daily_activity_summary for daily goals and totals, get_heart_rate and get_heart_rate_by_date_range for heart-rate data, and get_nutrition/get_food_log for nutrition details. You can also pull activity time series and azm (Active Zone Minutes) data, and fetch profile details. Use these tools by issuing structured requests to the MCP endpoint, enabling your AI assistant to answer questions like “Show my sleep patterns this week” or “What were my workouts last month?”

How to install

Prerequisites:

  • Node.js installed on your machine (recommended version per the project’s requirements)
  • Internet access to install dependencies and fetch Fitbit OAuth credentials

Option A: Install from npm (recommended for end users)

  1. Obtain Fitbit API credentials (see Fitbit developer portal). Create a Personal OAuth 2.0 app and note the Client ID and Client Secret.
  2. Install the MCP Fitbit package globally:
npm install -g mcp-fitbit
  1. Create a .env file in the package directory as directed by the tool when you first run it, and add:
FITBIT_CLIENT_ID=your_client_id_here
FITBIT_CLIENT_SECRET=your_client_secret_here
  1. Run the server:
mcp-fitbit

Option B: Development setup from source

  1. Get Fitbit API credentials as above.
  2. Clone and install:
git clone https://github.com/TheDigitalNinja/mcp-fitbit
cd mcp-fitbit
npm install
  1. Create a .env file:
FITBIT_CLIENT_ID=your_client_id_here
FITBIT_CLIENT_SECRET=your_client_secret_here
  1. Build or run in dev mode as appropriate per the repository (e.g., npm run dev).

Both options expose a local MCP Inspector UI at http://localhost:5173 for interactive testing and OAuth flows.

Additional notes

Tips and common issues:

  • Ensure your Fitbit OAuth credentials are correctly configured and that the redirect/callback URL matches what Fitbit expects (e.g., http://localhost:3000/callback).
  • When configuring Claude Desktop, you can add the fitbit MCP server with command mcp-fitbit or point to a local dev instance via Node if you're developing locally.
  • The availability and behavior of individual tools may depend on the OAuth scope granted by your Fitbit app. If a tool returns missing data, verify the scope and that the access token is valid.
  • If you’re running locally, the MCP Inspector UI is a helpful environment for testing tool calls and OAuth flows before integrating into production workflows.
  • Environment variables in the config are placeholders; replace with your actual Fitbit credentials before running in a production setting.

Related MCP Servers

Sponsor this space

Reach thousands of developers