Can Claude read my entire Apple Health export?

Not effectively. Apple Health exports can be hundreds of MB to 1GB+. The better workflow is to query specific metrics and date ranges via a local API, then send the compact results to Claude for analysis.

What is the base URL for the Apple Health local API?

By default the API runs locally at http://127.0.0.1:8765.

How do I authenticate to the API?

Most routes require the header X-Health-Analyzer-Token: . The token file path is exposed by GET /status.

Can Claude call the API automatically?

Yes—if your Claude setup supports tools/function calling. Otherwise you can paste the API JSON response into Claude manually.

Claude + Local API

Apple Health + Claude (via a Local API)

Claude can reason extremely well—but it can’t (and shouldn’t) ingest your full Apple Health export. Query the exact metric slice locally, then send a compact result for analysis.

Export → AirDrop → Analyze with Claude, ChatGPT, or any AI.

Prerequisite: you must install the Mac app (it hosts the local API on your machine). No Mac app → no API.

Download Mac App Setup guide

Base URL: http://127.0.0.1:8765
Auth header (most routes): X-Health-Analyzer-Token: <token>
Token file path: returned by GET /status

Docs: Apple Health Local API · Hub: Apple Health + AI

What Claude is great at (once it has the right slice)

Weekly reviews: “What changed vs last week?”
Trend detection: resting HR drift, sleep consistency, training load shifts
Hypotheses + next queries: what to pull next to confirm an idea
Plans: turn findings into an actionable 7–14 day experiment

Quickstart

Confirm the app is running: GET http://127.0.0.1:8765/status
Read the token file path from /status, load the token
Call a summary endpoint with X-Health-Analyzer-Token
Paste the JSON into Claude (or wire it up via tools/function calling)

Example Claude prompts (copy/paste)

Prompt 1 — Weekly review

Analyze the JSON below for the last 30 days.

1) 5-bullet executive summary
2) Biggest week-over-week deltas
3) 2 hypotheses for what caused the change
4) The next 3 API queries I should run (endpoint + date range)

Return concise bullet points.

Prompt 2 — Recovery signal check

Based on these heart-rate trends and sleep summaries:

- Do you see signs of overreaching, stress, or illness?
- What signals are most predictive here?
- What would you track next to increase confidence?

Be explicit about confidence and alternative explanations.

Example API calls (minimal data in)

Discover metrics:

curl -s \
  -H "X-Health-Analyzer-Token: $TOKEN" \
  "http://127.0.0.1:8765/metrics"

Sleep summary (date range):

curl -s \
  -H "X-Health-Analyzer-Token: $TOKEN" \
  "http://127.0.0.1:8765/sleep/summary?start=2026-02-25&end=2026-03-27"

Heart-rate trends (date range):

curl -s \
  -H "X-Health-Analyzer-Token: $TOKEN" \
  "http://127.0.0.1:8765/heart-rate/trends?start=2026-02-25&end=2026-03-27"

Privacy & security notes

API runs locally on 127.0.0.1
Most endpoints require X-Health-Analyzer-Token
Share only the summary/trend payload you want Claude to analyze

Compatibility routes

Compatibility endpoints also exist under /health/*, for example /health/sleep and /health/workouts.

Ready to query your data?

Start with the Local API docs, then build your Claude workflow.

Local API Docs Next: OpenAI / ChatGPT