briangaoo

briangaoo / whoop-mcp

Public

Model Context Protocol server giving Claude (or any MCP client) full read + write access to your Whoop fitness data via the private reverse-engineered iOS API. 47 tools: recovery, sleep, strain, HRV trends, Strength Trainer, journal, Whoop Coach, and smart alarm. TypeScript + zod, auto-refresh Cognito auth.

anthropic claude fitness mcp model-context-protocol
10
0
69% credibility
Found May 29, 2026 at 12 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

Whoop-MCP is an open-source tool that connects your Whoop fitness tracker to AI assistants like Claude. It unlocks 48 different ways to interact with your Whoop data - from reading detailed recovery scores and sleep analysis to logging strength training workouts and journal entries. The tool runs locally on your computer or can be deployed to a server, giving you full control over your credentials. It includes safety features like a 'preview first' system for any changes, so you always see what will happen before it happens. The project is transparent about using Whoop's internal app interface rather than their official developer API, and includes clear disclaimers about the trade-offs involved.

How It Works

1
💪 You want deeper insights from your Whoop

You've been using Whoop to track your fitness, but the app only shows you pieces of the picture. You want to ask an AI assistant questions about your recovery, sleep, and workouts in natural language.

2
🔌 You connect Whoop to your AI assistant

You run a simple setup command that connects your Whoop account to Claude or another AI assistant. The setup walks you through logging in with your Whoop email and password.

3
✨ Your AI assistant gains superpowers

Suddenly your AI can see everything Whoop tracks: your recovery score, detailed sleep stages, strength training history, journal entries, stress levels, and more. It can even write data back, like logging workouts or updating your profile.

4
You ask questions your way
📊
Read mode

Ask 'How did I recover today?' or 'Show me my sleep trends over the past month' and get detailed, structured answers.

✏️
Write mode

Log workouts, update your journal, set smart alarms, or configure your heart rate zones - with a preview step so nothing happens without your approval.

5
🔒 Your data stays private

Everything happens on your own computer or your own server. Your Whoop credentials never leave your control, and there's no tracking or analytics involved.

🎉 You get a personal fitness analyst

Now you have a knowledgeable assistant that understands your Whoop data inside and out - ready to answer questions, spot patterns, and help you log your activities whenever you need.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 12 to 10 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is whoop-mcp?

whoop-mcp is a Model Context Protocol server that gives AI assistants like Claude full read and write access to your Whoop fitness data. Instead of using Whoop's limited public OAuth API (13 read-only endpoints), it wraps Whoop's private iOS API—the same one the Whoop app uses internally. The result: 48 tools covering recovery scores, sleep hypnograms, HRV trends, Strength Trainer logging, journal entries with 308 behavior types, Whoop Coach chat, smart alarms, and more. TypeScript with zod validation ensures responses are structured and predictable. A write-safety harness requires explicit confirmation before any mutation, so a hallucinating AI cannot silently create garbage data on your account.

Why is it gaining traction?

The public Whoop API is read-only and missing nearly everything Whoop tracks. This project unlocks the full feature set—including programmatic writes, which no official path exists for. The `whoop-mcp cloud` command deploys the server to Fly, Railway, or Koyeb in one guided flow and connects it to Claude across web, desktop, and mobile via OAuth. Built-in catalogs (372 exercises, 308 journal behaviors, 203 sports) are bundled at build time, not fetched at runtime, keeping token usage predictable. When Whoop inevitably changes response shapes, zod validation surfaces drift as explicit errors rather than silent data corruption.

Who should use this?

Whoop members who want AI assistants to analyze their fitness trends, automate journal logging, or manage Strength Trainer workouts without opening the app. Developers building Whoop integrations will find this the fastest path to the full API surface. Fitness data nerds who want to correlate recovery metrics with behavior patterns (caffeine, sleep environment, supplements) across Whoop's 308 tracked behaviors. Not for casual users—setup requires Node 24, a deployed server for cross-device sync, and acceptance that this uses a reverse-engineered API with no official support.

Verdict

This is a serious piece of engineering solving a real gap: Whoop's public API is nearly useless for power users. The documentation is thorough, the write-safety design is thoughtful, and the CLI makes deployment approachable. However, with 10 stars and a 0.699% credibility score, this is early-stage software from a solo maintainer. Whoop could break the iOS API at any update. Use it if you understand the risk and need the features—check the public OAuth API first if your use case is read-only.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.