Fitbit MCP is an unofficial Model Context Protocol implementation that lets AI assistants query and analyze a user's Fitbit health and fitness data through Fitbit's public API. It exposes a suite of MCP tools such as getUserProfile, getActivities, getSleepLogs, getHeartRate, getSteps, getBodyMeasurements, getFoodLogs, getWaterLogs, getLifetimeStats, getUserSettings, getFloorsClimbed, getDistance, getCalories, getActiveZoneMinutes, getDevices, and getBadges. Access is token-based, using a Fitbit access token obtained via the OAuth 2.0 flow, provided either via the FITBIT_ACCESS_TOKEN environment variable or as a command-line argument. Most tools accept optional parameters like date (default today) and period (1d, 7d, 30d, 1w, 1m). The README describes configuring and running the MCP with JSON configuration for AI frameworks, and includes setup and release steps for npm. A disclaimer notes that this is an unofficial integration not affiliated with Fitbit. The repository provides development commands (build, dev) and GitHub Actions workflows for automated validation.