Python Google Health API Client Library

An asynchronous, type-safe Python client library for the Google Health API (health.googleapis.com/v4). This library is designed to help developers migrate from the legacy Fitbit Web API to the Google Health API.

It exposes Google Health data types as clean, type-annotated properties on the client (e.g., api.steps and api.heart_rate).

Features

• Asynchronous: Built on aiohttp for non-blocking asynchronous requests.
• Type-safe: Leverages Python dataclasses and mashumaro for seamless JSON serialization and deserialization, matching the exact Google Health API schemas.
• AIP-160 Filter Support: Built-in helper to automatically translate timestamp queries into Google API filter expressions.
• Auto-pagination: Paginated results wrap list endpoints and provide asynchronous iterators to traverse pages easily.

Supported Data Types

Currently supported data types:

• Steps (api.steps)
• Heart Rate (api.heart_rate)
• Sleep (api.sleep)
• Distance (api.distance)
• Basal Energy Burned (api.basal_energy_burned)
• Active Energy Burned (api.active_energy_burned)
• Total Calories (api.total_calories)
• VO2 Max (api.vo2_max)
• Weight (api.weight)
• Floors (api.floors)
• Hydration Log (Water) (api.hydration_log)
• Nutrition Log (Food) (api.nutrition_log)
• Daily Resting Heart Rate (api.daily_resting_heart_rate)

Installation

Install the package from PyPI using uv:

uv pip install google-health-api

Quickstart

import asyncio
import aiohttp
from google_health_api.api import GoogleHealthApi
from google_health_api.auth import AbstractAuth

class SimpleAuth(AbstractAuth):
    def __init__(self, websession: aiohttp.ClientSession, access_token: str) -> None:
        super().__init__(websession)
        self._access_token = access_token

    async def async_get_access_token(self) -> str:
        return self._access_token

async def main():
    async with aiohttp.ClientSession() as session:
        # Initialize authorization wrapper with your Google OAuth 2.0 access token
        auth = SimpleAuth(session, "YOUR_ACCESS_TOKEN")
        api = GoogleHealthApi(auth)

        # 1. Fetch steps from the last 24 hours
        from datetime import datetime, timezone, timedelta
        end_time = datetime.now(timezone.utc)
        start_time = end_time - timedelta(days=1)

        result = await api.steps.list(start_time=start_time, end_time=end_time)
        print("Steps:")
        for point in result.data_points:
            print(f"  Count: {point.data.count} steps")
            print(f"  Time: {point.data.start_time} to {point.data.end_time}")

        # 2. Fetch heart rate records
        hr_result = await api.heart_rate.list(start_time=start_time, end_time=end_time)
        print("Heart Rates:")
        for point in hr_result.data_points:
            print(f"  BPM: {point.data.bpm}")

if __name__ == "__main__":
    asyncio.run(main())

Development and Testing

Verify code quality and type safety:

./script/lint

Run mock tests:

./script/test

Verify with a live account using the browser-based OAuth 2.0 Installed App Flow:

1. Create an OAuth client ID for a "Desktop application" in the Google Cloud Console.
2. Download the JSON key file, rename it to client_secret.json, and place it in the project root.
3. Authenticate and query via the command-line interface tool:

   # Log in via your web browser (stores credentials in token.json)
   google-health-cli login
   
   # List step data points
   google-health-cli steps list --days 7 --limit 5
   
   # List heart rate data points
   google-health-cli heart-rate list --days 7 --limit 5
   
   # Retrieve user profile details
   google-health-cli profile get

Command-Line Interface (CLI)

The package installs a real binary executable google-health-cli designed with Agent DX (AI Agent Developer Experience) and Human DX principles.

Key Capabilities

• Dynamic Schema Introspection: Let agents query request/response layouts at runtime.

  google-health-cli schema
  google-health-cli schema steps.list

• Raw JSON Payloads: Pass structured queries or write payloads directly to bypass flat CLI argument limits.

  # Query filtering via --params
  google-health-cli --params '{"pageSize": 5, "startTime": "2026-06-26T00:00:00Z"}' steps list
  
  # Mutation via --json
  google-health-cli --json '{"steps": {"count": 100, ...}}' steps create

• Context Token Discipline: Use fields/masking filters to prevent bloating agent reasoning limits.

  google-health-cli --fields "dataPoints(steps(count,interval))" steps list

• Safety Rails & Dry-Runs: Validate mutating requests locally before hitting the API.

  google-health-cli --dry-run --json '{"steps": {"count": 100, ...}}' steps create

• Headless Integration: Autodetects environment variables for credentials in headless agent contexts.

  export GOOGLE_HEALTH_CLI_TOKEN="YOUR_OAUTH_TOKEN"
  google-health-cli profile get