The app shows each metric on its own screen; here every stream arrives in one conversation, so questions like how last week's sleep score tracked with training load become answerable. This is one person's project, offered as-is with no guarantee of uptime, and it holds an OAuth token to your health data — tokens are encrypted at rest with AES-256-GCM and the code is public for that reason. If that trade does not sit right, the same repository documents running your own instance.
A hosted MCP server that connects to your Withings account over OAuth and exposes your health history as tools an assistant can query.
- get_sleep_summary — duration, light/deep/REM stages, heart rate, breathing and sleep score
- get_hrv — minute-level RMSSD, SDNN and quality scores captured during sleep
- get_activity and get_intraday_activity — daily steps, distance, calories, elevation, and higher-frequency data through the day
- get_workouts — workout summaries with heart rate zones and metrics
- get_measures — weight, body composition, blood pressure, heart rate, temperature, VO2 max and more
- list_heart_records and get_heart_signal — your ECG recordings and their waveform data
- list_stetho_records and get_stetho_signal — stethoscope recordings, if you have a BPM Core
- get_user_devices and get_user_goals — the devices on the account and the step, sleep and weight goals set on it
A Withings account with connected devices, authorized over OAuth. Add https://withings-mcp.com/mcp as a custom connector and approve the permissions in the browser — nothing to install. Access is revocable from your Withings account settings at any time. Self-hosting is documented as well and needs Bun 1.1 or higher, a Withings developer application, and a Supabase project.