Skip to content
MCP server

Health Access MCP

Read Apple Watch workouts and vitals from local files, and send AI-written plans back

Runs locallyStatus and history
Install
Pricing
Open source
Platforms
iOS, CLI, macOS
Runs
Locally
Version
1.0.4
Listing updated
3 days ago

About Health Access MCP

Health Access MCP is an open-source MCP server for health and science. It runs locally on your machine and is started by your AI client.

Apple does not let a desktop AI read Watch data, so this works through a companion iOS app that writes each finished workout and daily vitals as JSON into a folder you choose, such as iCloud Drive, Dropbox or a NAS. The MCP server runs on your computer with npx, finds that folder on its own and reads it over stdio. There is no account and no server; the app's settings screen shows what left the phone, which stays empty.

For each session the files include distance, time, heart rate every few seconds, intervals with names and, for swims, strokes and SWOLF. The AI can list and open workouts, read resting heart rate, HRV, sleep, weight, steps and VO2max, and get weekly summaries. It can also write a training plan: it is validated, saved to a plans folder, appears in the app and goes onto the Watch with one tap or automatically, with step labels shown on the wrist on iOS 18.

The site lists limits that come from Apple: data exports only while the phone is unlocked, and pool length cannot be set by a plan. It needs Node.js 18 or newer. The server and command line are MIT licensed. ChatGPT works through the files, not the MCP, since this server uses stdio.

Install

Run in a terminal

claude mcp add healthaccess-mcp -- npx -y healthaccess-mcp

Add to claude_desktop_config.json

{
  "mcpServers": {
    "healthaccess-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "healthaccess-mcp"
      ]
    }
  }
}
Install in CursorOpens Cursor, which asks you to confirm. Or add it by hand:

Add to ~/.cursor/mcp.json

{
  "mcpServers": {
    "healthaccess-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "healthaccess-mcp"
      ]
    }
  }
}
Install in VS CodeOpens VS Code, which asks you to confirm. Or add it by hand:

Add to .vscode/mcp.json

{
  "servers": {
    "healthaccess-mcp": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "healthaccess-mcp"
      ]
    }
  }
}

Add to ~/.codeium/windsurf/mcp_config.json

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

Add to your client's MCP settings (most use this shape)

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

Tools this server provides (8)

Taken from the project documentation; the running server may differ.

list_workouts
Compact rows of sessions, newest first.
get_workout
One session in full with intervals, pace, heart rate and lengths.
get_vitals
Resting heart rate, HRV, sleep, weight, steps and VO2max by day.
training_summary
Weekly volume, sessions, easy-zone share, effort and cardiac cost trend.
rate_workout
Stores the effort rating you give.
push_plan
Validates and writes a workout plan for the Watch.
get_plan_status
Shows whether a plan reached the Watch.
list_plans
Lists plans and their states.

What it does

Local file export
Writes workouts and vitals as JSON into a folder you pick.
Workout detail
Returns intervals, pace, heart rate, zone time and lengths for a session.
Vitals and trends
Reads resting heart rate, HRV, sleep, weight, steps, VO2max and weekly summaries.
Plans to the Watch
push_plan validates a session and sends it to the Watch via the app.
Effort ratings
Stores an effort score you tell the AI, never guessed.
Briefing file
A one-page briefing.md works with any AI that can read a file.

Good for

  • Ask for a weekly check-in on training volume
  • Explain your last swim session
  • Check whether you are overdoing it from effort and heart rate
  • Have an AI plan your next session and put it on the Watch
  • Prepare a summary for a coach or doctor

Pricing

Open source

The iOS app is free; server and CLI are MIT licensed

Prices change. Check healthaccessmcp.com before you buy.

Health Access MCP FAQ

Does it need an account?

No. There is no account and no server; data stays in the folder you pick.

Does it work with ChatGPT?

Through the files, yes; through the MCP not today, because it uses stdio.

What are the Apple limits?

Health data can be read only while the phone is unlocked, and plans cannot set the pool length.

What is the licence?

The page says the server and command line are MIT.

Is Health Access MCP free?

Yes. Health Access MCP is open source, so you can use and self-host it for free.

How do I add Health Access MCP to Claude Code?

Run this command in your terminal: claude mcp add healthaccess-mcp -- npx -y healthaccess-mcp

What are the best alternatives to Health Access MCP?

The closest alternatives to Health Access MCP are Tooluniverse, Oya Browser, Later Stages, MedXpert 标准检索连接器 and MedXpert 全球法规连接器 (reg-connector).

Alternatives to Health Access MCP

  1. Tooluniverse

    MCP server

    MCP server giving AI agents access to thousands of scientific research tools

    Runs locally
    • Open source
    • 1.8K GitHub stars
    • Health and science
  2. Oya Browser

    MCP server

    HIPAA-oriented agentic browser with replayable playbooks, reachable over MCP

    • 32 tools
    • Remote
    • Freemium
    • 349 GitHub stars
    • Browser automation
  3. Later Stages

    MCP server

    California senior living and care provider records from state licensing and Medicare

    • 6 tools
    • Remote
    • Free
    • Health and science
  4. Read-only local index of medical-device standards with official links, no standard text

    Runs locally
    • 4 tools
    • Free
    • Legal and government
  5. Splyt

    MCP server

    Let Claude or ChatGPT read your Splyt workouts, records, plan, habits and sleep

    • Remote
    • Free
    • Health and science