Huawei MCP Connector
Provides read-only access to Huawei Health data, including workouts, sleep records, resting heart rate, HRV, running ability, training summaries, weekly trends, and personal bests.
This server does not currently have a deployable release on Glama. Follow the repository's setup instructions to install and configure it for your MCP client, including any required credentials and local dependencies.
Huawei MCP Connector
A local MCP server that lets an AI assistant such as Claude read your Huawei Health data: workouts, sleep, resting heart rate, HRV, SpO2, stress, emotion, running ability and personal bests.
Read-only. It only queries data. It never writes.
Logs in for you. You give it your Huawei ID account and password once. It signs in, asks you for the one-time SMS code the first time, then refreshes the token by itself from then on — no more codes, no hourly copy-paste.
Single user, your machine. Everything stays local. Runs only via Docker Compose.
How login works
The server drives the real Huawei login in a headless browser it keeps inside the container:
First run: it enters your account + password, Huawei sends an SMS code, you give the code to the assistant, and it marks the browser as a trusted device.
After that: when the ~1-hour token expires, it logs in again silently with just the account and password (trusted device = no code). You only see a code again if Huawei drops the trust.
The headless browser is used only to obtain the token; data requests use plain HTTPS with it.
Related MCP server: health-mcp
Tools
Tool | Purpose |
| Whether a valid session is held |
| Begin login; sends the SMS (or logs in silently if already trusted) |
| Finish the first login with the SMS code; marks the device trusted |
| Recent workouts (distance, HR, pace, VO2max, training load) |
| Raw samples for one workout (HR, speed, cadence, GPS, altitude, power) |
| Min/max/avg per sensor stream for one workout |
| Headline metrics for one workout (duration, HR, speed, pace, distance) |
| Totals for the last N days, also by activity type |
| Last N days vs the N before |
| Weekly totals for the last N weeks |
| Week-over-week changes |
| Nightly sleep stages, score, efficiency, wake-ups |
| Daily resting heart rate |
| Nightly HRV |
| Daily blood-oxygen (SpO2) |
| Daily stress score |
| Happy / calm / unhappy emotion readings |
| Running ability, fatigue, predicted race times |
| Personal records |
| Server liveness check |
Setup
1. Configure your account
Create a .env file next to docker-compose.yml (it is git-ignored):
HUAWEI_ACCOUNT=your_huawei_id_phone_or_email
HUAWEI_PASSWORD=your_password
HUAWEI_TZ=Asia/Ho_Chi_Minh # your timezone, for daily heart-rate/HRV windows
BIND=127.0.0.1 # this machine only2. Start it
docker compose up -d --buildFirst build pulls the Playwright browser image (~1.8 GB) and takes a few minutes. The server then
listens on http://localhost:8000/mcp.
3. Connect Claude Code
claude mcp add --scope local --transport http huawei http://localhost:8000/mcpNo token or header needed — the server authenticates itself.
4. First login
In Claude, ask it to start login. Claude calls start_login, Huawei texts you a code, you tell
Claude the code, and Claude calls submit_sms_code. Done — from now on it refreshes silently.
Then ask things like "How did I sleep this week?" or "What's my 7-day running volume vs last week?"
Configuration
Set these in .env:
Variable | Default | Purpose |
| — (required) | Huawei ID phone or email |
| — (required) | Huawei ID password |
|
| IANA timezone for daily resting-HR/HRV windows |
|
| Host interface to publish on |
|
| Host port |
The trusted browser profile is kept in a Docker volume (huawei-profile), so it survives restarts and
you don't re-enter a code after docker compose up.
Security
Your Huawei account and password sit in
.envon your machine and in the container's environment. They control your whole Huawei ID, so keep.envprivate; it is git-ignored.The trusted-device cookie lives in the
huawei-profilevolume. Treat it like a login session.The server binds to
127.0.0.1by default. To reach it from elsewhere, deploy to a server withBIND=0.0.0.0behind HTTPS (e.g. Caddy or nginx) — never expose it over plain HTTP.
Caveats
Rate limiting: Huawei throttles both SMS codes and login attempts if you make many in a short time. If
start_loginsays an SMS was sent but none arrives, or a tool reports "Huawei is rate-limiting logins", wait a while (up to a few hours) and try again. In normal use this never happens — a trusted device refreshes silently without new codes.Captcha: Huawei can demand a captcha during login (risk-based; rare on a trusted device). If it does,
start_loginreturnscaptcha_requiredand you'll need to log in once in a normal browser.Region: detected automatically from your account at login (Asia / Europe / China).
License
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
This server cannot be deployed
Maintenance
Related MCP Connectors
Garmin watch data for AI assistants: activities, sleep, HRV, recovery. Never writes to Garmin.
Your Garmin, Oura, WHOOP, Hevy and 35 more fitness apps, queryable from your own AI.
Connect your wearables, rings and training apps, then ask your AI about your own health data.
Garmin data in Claude & ChatGPT via the Garmin Health API. OAuth sign-in, no password sharing.
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceEnables AI assistants to access your Oura Ring health data through OAuth2-authenticated API calls.9MIT
- AlicenseNot gradedqualityBmaintenanceEnables AI agents to query personal health data collected by any wearable or phone fitness app through Android Health Connect or Apple Health, exposing read-only tools for metrics like heart rate, sleep stages, VO2 max, and body composition. It can run locally over USB with nothing stored, or be self-hosted on Cloudflare Workers with D1 to retain history beyond Health Connect's 30-day limit.20 npmMIT
- AlicenseNot gradedqualityBmaintenanceEnables AI assistants to read Fitbit/Google Health data (steps, sleep, heart rate, SpO2, workouts) and log weight, food, water, and exercises via natural language, self-hosted on Cloudflare Workers.5MIT
- AlicenseNot gradedqualityBmaintenanceEnables AI assistants such as Claude, ChatGPT and Gemini to read wearable health data — daily activity, sleep stages, heart rate, SpO2, HRV, workouts, weight and nutrition — through a self-hosted, OAuth-protected Cloudflare Worker endpoint. It also lets users log weight, body fat, water, food, exercises, mood and symptoms by voice or chat, while leaving sensor-recorded data read-only.MIT