Sponsored by Deepsite.site

Whoop Mcp Unofficial

Created By
davidmosiah2 days ago
Unofficial local-first MCP server for WHOOP — recovery, strain, sleep stages, HRV, RHR and workouts. Privacy modes (summary/structured/raw), SQLite cache and doctor CLI for setup. Works with Claude Desktop, Cursor, Windsurf, Hermes, OpenClaw and any MCP-compatible client. Install: `npx -y whoop-mcp-unofficial setup`. Built for self-quantified humans building AI workflows around personal health.
Overview

whoop-mcp-server

MCP Compatible License: MIT TypeScript Provider: WHOOP npm version GitHub stars npm downloads CI Delx Wellness Agent-ready MCP

Local-first MCP server that connects AI agents to your WHOOP recovery, sleep, strain and HRV data.

Unofficial project. Not affiliated with, endorsed by or supported by WHOOP, Inc. WHOOP is a trademark of its respective owner. Use this only with your own WHOOP account and in line with WHOOP's Developer Terms.

Built by David Mosiah for people who use Claude, Cursor, Hermes, OpenClaw or other MCP-compatible agents to think about training, sleep and recovery — without copy-pasting numbers from the WHOOP app.

Part of Delx Wellness, a registry of local-first wellness MCP connectors.

If this connector helps your agent workflow, please star the repo. Stars make the project easier for other AI builders to discover and help Delx keep shipping local-first wellness infrastructure.

Why this exists

WHOOP gives you rich physiology — recovery score, HRV, sleep stages, strain — but it lives behind an OAuth API and a closed app. Bringing it into your AI agent today means writing the OAuth dance yourself, storing tokens safely, normalizing responses and handling pagination.

This package does all of that locally, exposes WHOOP through the Model Context Protocol, and lets any MCP-compatible agent read your WHOOP context with one config snippet. Tokens never leave your machine.

Setup in 60 seconds

You'll need a WHOOP Developer app (create one here) with redirect URI http://127.0.0.1:3000/callback.

npx -y whoop-mcp-unofficial setup    # interactive: paste client id + secret
npx -y whoop-mcp-unofficial auth     # opens browser, captures the OAuth code
npx -y whoop-mcp-unofficial doctor   # verifies you're ready

Then add this to your MCP client config:

{
  "mcpServers": {
    "whoop": {
      "command": "npx",
      "args": ["-y", "whoop-mcp-unofficial"]
    }
  }
}

For Claude Desktop, run setup --client claude and the snippet is written for you.

Try it with your agent

Three things to ask first:

Use whoop_connection_status to check setup, then run whoop_daily_summary.
Give me a 5-line operating brief for today.
Call whoop_weekly_summary with response_format=json. Identify the top
bottleneck and give me a sleep + training plan for next week.
Use the whoop_daily_performance_coach prompt. Focus on whether I should train
hard today.

Data availability

This package uses the official WHOOP OAuth API (v2). It does not access raw device sensor streams.

DataAvailableNotes
Recovery score, HRV, RHR, SpO2, skin tempWhen WHOOP returns a scored recovery
Sleep sessions + stages + performanceAll scored sleep records
Cycles + day strain + kilojoulesPhysiological cycles
Workouts + sport + heart-rate zonesAll recorded workouts
Profile + body measurementsHeight, weight, max HR
Continuous heart-rate / device telemetryNot exposed by WHOOP's public API
Live BLE heart-rate listeningThis package is not a Bluetooth listener

When this README says raw, it means the upstream WHOOP API JSON for a supported endpoint — not raw sensor samples.

Tools

Start with these:

  • whoop_connection_status — verify local setup before calling WHOOP
  • whoop_daily_summary — readiness, sleep, load and action candidates for today
  • whoop_weekly_summary — scorecard, comparison vs prior week, next-week plan

Auth & diagnostics

  • whoop_capabilities, whoop_agent_manifest, whoop_privacy_audit, whoop_cache_status
  • whoop_get_auth_url, whoop_exchange_code, whoop_revoke_access

Profile

  • whoop_get_profile, whoop_get_body_measurements

Collections (paginated, with start/end filters and privacy-mode override)

  • whoop_list_recoveries, whoop_list_sleeps, whoop_list_cycles, whoop_list_workouts

Common collection params: start, end, limit (max 25), next_token, all_pages, max_pages, response_format (markdown/json), privacy_mode (summary/structured/raw).

Single records by id

  • whoop_get_cycle, whoop_get_sleep, whoop_get_workout
  • whoop_get_cycle_sleep, whoop_get_cycle_recovery

Prompts

  • whoop_daily_performance_coach — practical daily plan from today's signals
  • whoop_weekly_training_review — week comparison + next-week plan
  • whoop_sleep_recovery_investigator — investigate sleep ↔ recovery patterns

Each accepts timezone (IANA, default UTC).

Resources

  • whoop://capabilities
  • whoop://summary/daily, whoop://summary/weekly
  • whoop://latest/recovery, whoop://latest/sleep, whoop://latest/cycle

Privacy & security

  • OAuth tokens are stored in ~/.whoop-mcp/tokens.json with 0600 permissions and are never returned by tools.
  • Refresh-token rotation uses a lock file to avoid concurrent refresh races.
  • whoop_revoke_access is the only destructive tool — it deletes local tokens and revokes the grant.
  • WHOOP_PRIVACY_MODE defaults to structured. Raw WHOOP API payloads are opt-in via raw mode or per-call override.
  • The MCP client never sees access or refresh tokens.
  • This is not medical advice. The server exposes user-authorized data for personal AI workflows, not diagnosis or treatment.

Configuration

setup writes most of these into ~/.whoop-mcp/config.json (0600). Manual env override is supported:

WHOOP_CLIENT_ID=WHOOP_CLIENT_SECRET=WHOOP_REDIRECT_URI=http://127.0.0.1:3000/callback

# Optional
WHOOP_SCOPES="read:recovery read:cycles read:workout read:sleep read:profile read:body_measurement"
WHOOP_PRIVACY_MODE=structured        # summary | structured | raw
WHOOP_CACHE=sqlite                   # optional read-through cache
WHOOP_TOKEN_PATH=~/.whoop-mcp/tokens.json
WHOOP_CACHE_PATH=~/.whoop-mcp/cache.sqlite

Hermes / remote setup

npx -y whoop-mcp-unofficial setup --client hermes --no-auth
npx -y whoop-mcp-unofficial auth                       # run locally if browser auth is needed
npx -y whoop-mcp-unofficial doctor --client hermes
hermes mcp test whoop

After Hermes config changes, use /reload-mcp or hermes mcp test whoop. Don't restart the gateway for normal data access.

If browser OAuth has to happen on a different machine than Hermes, run auth locally and copy ~/.whoop-mcp/tokens.json to the server with chmod 600.

Requirements

  • Node.js 20+
  • A WHOOP Developer app with redirect URI http://127.0.0.1:3000/callback

Default OAuth scopes:

read:recovery read:cycles read:workout read:sleep read:profile read:body_measurement

Development

git clone https://github.com/davidmosiah/whoop-mcp.git
cd whoop-mcp
npm install
npm test
npm run build

Test with MCP Inspector:

npx @modelcontextprotocol/inspector node dist/index.js

Optional local HTTP transport:

WHOOP_MCP_TRANSPORT=http WHOOP_MCP_PORT=3000 node dist/index.js
curl http://127.0.0.1:3000/health

Docs

License

MIT — see LICENSE.

Disclaimer

This software is provided as-is. It is not a medical device, does not provide medical advice, and should not be used for diagnosis or treatment. Always consult qualified professionals for medical concerns.

Server Config

{
  "mcpServers": {
    "whoop": {
      "command": "npx",
      "args": [
        "-y",
        "whoop-mcp-unofficial"
      ]
    }
  }
}
Recommend Servers
TraeBuild with Free GPT-4.1 & Claude 3.7. Fully MCP-Ready.
Tavily Mcp
MiniMax MCPOfficial MiniMax Model Context Protocol (MCP) server that enables interaction with powerful Text to Speech, image generation and video generation APIs.
Amap Maps高德地图官方 MCP Server
WindsurfThe new purpose-built IDE to harness magic
Jina AI MCP ToolsA Model Context Protocol (MCP) server that integrates with Jina AI Search Foundation APIs.
MCP AdvisorMCP Advisor & Installation - Use the right MCP server for your needs
AiimagemultistyleA Model Context Protocol (MCP) server for image generation and manipulation using fal.ai's Stable Diffusion model.
Playwright McpPlaywright MCP server
BlenderBlenderMCP connects Blender to Claude AI through the Model Context Protocol (MCP), allowing Claude to directly interact with and control Blender. This integration enables prompt assisted 3D modeling, scene creation, and manipulation.
Serper MCP ServerA Serper MCP Server
Y GuiA web-based graphical interface for AI chat interactions with support for multiple AI models and MCP (Model Context Protocol) servers.
Howtocook Mcp基于Anduin2017 / HowToCook (程序员在家做饭指南)的mcp server,帮你推荐菜谱、规划膳食,解决“今天吃什么“的世纪难题; Based on Anduin2017/HowToCook (Programmer's Guide to Cooking at Home), MCP Server helps you recommend recipes, plan meals, and solve the century old problem of "what to eat today"
Baidu Map百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
RedisA Model Context Protocol server that provides access to Redis databases. This server enables LLMs to interact with Redis key-value stores through a set of standardized tools.
EdgeOne Pages MCPAn MCP service designed for deploying HTML content to EdgeOne Pages and obtaining an accessible public URL.
Zhipu Web SearchZhipu Web Search MCP Server is a search engine specifically designed for large models. It integrates four search engines, allowing users to flexibly compare and switch between them. Building upon the web crawling and ranking capabilities of traditional search engines, it enhances intent recognition capabilities, returning results more suitable for large model processing (such as webpage titles, URLs, summaries, site names, site icons, etc.). This helps AI applications achieve "dynamic knowledge acquisition" and "precise scenario adaptation" capabilities.
DeepChatYour AI Partner on Desktop
Visual Studio Code - Open Source ("Code - OSS")Visual Studio Code
CursorThe AI Code Editor
ChatWiseThe second fastest AI chatbot™