zayden16/whoop-mcp-server — explained in plain English
Analysis updated 2026-05-18
Ask Claude for your recovery score, sleep quality, or daily strain without opening the Whoop app.
Build a coaching or habit assistant that pulls real Whoop metrics into its answers.
Track average strain over a custom time window automatically.
Pull raw Whoop data, cycles, sleep stages, and workouts, into any MCP-compatible AI tool.
| zayden16/whoop-mcp-server | 42wim/fabio | 42wim/go-xmpp | |
|---|---|---|---|
| Stars | 0 | — | — |
| Language | Go | Go | Go |
| Last pushed | — | 2018-02-04 | 2020-01-24 |
| Maintenance | — | Dormant | Dormant |
| Setup difficulty | moderate | moderate | moderate |
| Complexity | 2/5 | 3/5 | 3/5 |
| Audience | developer | ops devops | developer |
Figures from each repo's GitHub metadata at analysis time.
Requires creating a Whoop developer app and completing a one-time OAuth login.
whoop-mcp-server connects an AI assistant like Claude to your Whoop fitness tracker data. It is built with the Model Context Protocol, or MCP, which is a way for tools like Claude Code or Claude Desktop to call out to small helper programs and get real information back instead of guessing. This project acts as that helper program for anyone who wears a Whoop band. It talks to the official Whoop API and exposes your health data as a set of tools an assistant can call: your profile, body measurements, daily cycles, recovery scores, sleep stages, and workouts. Each tool returns the same data shapes Whoop's own API provides, so nothing is simplified or hidden, all the detail Whoop tracks, such as heart rate, strain, and sleep stage durations, stays available to the model. There is also a built in tool for calculating your average strain over a chosen number of days. Setting it up takes a few steps. You first register a small app on Whoop's developer site to get a client ID and secret, then install the server itself, using Homebrew, a prebuilt binary, or building from source with Go. After that you run a one time authorization command that opens a browser so you can log into Whoop and grant access. The resulting login token is saved locally and refreshes itself automatically from then on. The final step is registering the server with Claude Code or Claude Desktop by pointing it at the installed program along with your Whoop credentials. Once connected, you can ask plain questions like what is my recovery score today, show my sleep for the past week, or what is my average strain over the last seven days, and the assistant will pull real answers from your Whoop account. The project is released under the MIT license and is written in Go.
An MCP server that lets AI assistants like Claude read your Whoop fitness tracker data, including recovery, sleep, strain, and workouts.
Mainly Go. The stack also includes Go, MCP, Whoop API.
Use freely for any purpose, including commercial use, as long as you keep the copyright notice.
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly developer.
This repo across BitVibe Labs
Don't trust strangers blindly. Verify against the repo.