Apple Health for the Quantified Self: Build a Personal Health Knowledge Base, HealthSave

Quantified self with Apple Health is the practice of exporting your own metrics into a place you control, then asking the questions you actually have. The hard part is not the export; the hard part is choosing a destination that fits the workflow you already keep. This page is the decision tree: which destination matches which kind of question, and how HealthSave wires into each.

Last reviewed 2026-09-20 · Educational guide

The minimal setup that works

You do not need a homelab. The smallest setup that is genuinely useful is this: install HealthSave free, open the Export tab, choose JSON, export the last 7 days, paste the file into an AI assistant (ChatGPT, Claude, or a local model), and ask one or two questions you actually care about.

That is it. No account, no cloud, no server, no Obsidian, no Postgres. The full Apple Health and AI guide walks through the prompts and the gotchas. If the small setup stops answering your questions, the rest of this page is for you.

Pick the destination that matches your workflow

The export format follows the destination. Most quantified-self setups fall into one of four families:

None of these are mutually exclusive. The same JSON file goes into Obsidian, an AI assistant, and a Postgres database at the same time; you do not have to choose. The point is to pick the one that matches the question you are actually trying to answer.

Three sample journeys

The sleep optimizer

Question: does the change I made last week actually move my HRV and sleep stages, or am I imagining it?

Setup: HealthSave exports JSON for the last 30 nights. A short script splits the file into per-night markdown in an Obsidian vault. The user opens the daily note each morning, writes a one-line entry about the change (caffeine cutoff, room temperature, supplement), and Dataview pulls the HRV and sleep stages onto the same screen as the note. After three weeks, an AI assistant reads the combined dataset and the daily notes and answers in plain English: yes, your HRV rose by ~7 ms on the nights you stopped caffeine by 2 PM.

What made it work: the data and the notes lived in the same place.

The runner in a training block

Question: how is the current training block affecting my resting heart rate, HRV, and workout paces?

Setup: HealthSave Pro exports JSON for the last 90 days, scoped to heart rate, HRV, and workouts. The user pastes the file into Claude once a week and asks three questions: is my resting HR trending up or down, which workouts in this block had the highest cardiac drift, and how does my HRV recovery compare to last block. The answers get pasted into a running log note in Obsidian.

What made it work: a narrow scope (three metrics, 90 days), a real cadence (weekly), and a destination that can answer follow-up questions.

The person who just wants to know if the medication is working

Question: has my resting heart rate actually changed since I started the new prescription?

Setup: HealthSave exports the last 90 days of heart rate and HRV as JSON. One question to an AI: "compare my resting heart rate in the 30 days before my medication start date versus the 30 days after; report the median, the spread, and whether the trend is statistically meaningful." Five-minute answer; no scripts.

What made it work: the smallest possible setup that answered the question. No homelab required.

The principle behind all three

Quantified self is not "collect everything and look at it later." It is "decide the question, pick the slice of data that answers it, put it somewhere you will actually look at it, and review on a cadence." Everything else is implementation.

The export is the cheap part. The discipline of looking at the numbers on a real schedule, and adjusting the question when the answer is uninteresting, is the practice.

What HealthSave is and is not in this picture

• HealthSave is the on-ramp. It reads Apple Health on your iPhone, with your permission, and writes a clean JSON or CSV file to wherever the iOS share sheet can reach. Free for the last 7 days; Pro unlocks the multi-year archive and background sync.
• HealthSave is not the destination. It does not draw charts, host your data, or generate insights. Those are the destinations' jobs.
• HealthSave is not a medical device. It is a tool for personal recordkeeping and self-investigation. For medical decisions, use a clinician.

The architecture in one diagram

iPhone (Apple Health)
      │
      │  read-only HealthKit access
      ▼
   HealthSave (on-device, no cloud)
      │
      │  JSON / CSV / PDF (iOS share sheet)
      ▼
   your destination
      ├─ Obsidian vault
      ├─ ChatGPT / Claude / local LLM
      ├─ Numbers / Excel / Sheets
      ├─ your Postgres or DuckDB
      ├─ Grafana or Home Assistant
      └─ a PDF in your doctor's inbox

Every arrow ends at a destination you control. The destination decides what the data looks like next.

Honest limits

• Apple Health is iOS only. If your wearable writes to Android Health Connect instead, this is the wrong starting point.
• Pro is required for the long archive. Free exports the last 7 days, which is enough to ask a question about last week. It is not enough for trend work; for that you want Pro ($24.99 one-time, Family Sharing, no subscription).
• Accuracy is your wearable's. Consumer sensors are not clinical instruments. HealthSave moves whatever your device recorded; it does not validate it.
• The discipline is yours. No app, including HealthSave, will look at the numbers for you.

Choosing your first destination

If you do not already keep notes, start with the AI path: paste a 7-day JSON into ChatGPT or Claude and ask one real question. If you already take daily notes, start with Obsidian. If you already run a homelab, start with the self-hosted path. If you have a doctor visit coming up, start with the PDF path. The right starting point is the one that matches a workflow you already keep.

HealthSave is not a medical device. It is for informational purposes only and does not diagnose, treat, cure, or prevent any disease or condition. Privacy claims of zero data collection apply to the iOS app; if you choose to send data to a self-hosted destination, that destination is yours to operate.

FAQ

What does quantified self mean with Apple Health?

Tracking your own metrics over time to answer questions you actually have. Apple Health records heart rate, HRV, sleep stages, workouts, weight, blood oxygen, and dozens more; the quantified-self practice is choosing the slice that matters, exporting it to a place you control, and reviewing it on a cadence that fits the question.

Where should I send my Apple Health data?

It depends on the workflow you already keep. Obsidian if you take daily notes. A spreadsheet if you want charts and pivot tables without code. An AI assistant if you want answers in plain English. A self-hosted database if you want a private archive you can query for years.

Do I need HealthSave Pro?

For the last 7 days, no. JSON and CSV export of the most recent week is free, no account. For a multi-year quantified-self archive (the only way most trends become visible), Pro is the unlock: a one-time $24.99 with Family Sharing, no subscription.

Can I do quantified self entirely on-device?

Yes. HealthSave reads Apple Health on the iPhone and writes the file to wherever you point the iOS share sheet. Nothing routes through a HealthSave cloud because there is not one. Move the file into Obsidian, an AI app, a local folder, or your own server: same data, your choice of destination.

How is this different from a dedicated health-tracking app?

Most dedicated apps lock you into their dashboard, their charts, and their idea of what to look at. A quantified-self setup built on Apple Health plus a destination you own lets you ask your own questions, keep the data in a format you control, and leave any tool at any time.

What is the smallest setup that is actually useful?

HealthSave free, export JSON for the last 7 days once a week, paste it into ChatGPT or Claude, and ask one or two questions you actually care about. That is enough to start. The bigger setups earn their complexity only when a smaller setup stops answering your questions.

What about Obsidian specifically?

Obsidian is one of the best destinations for a quantified-self setup because it is yours: a folder of markdown files on a disk you control. HealthSave exports JSON; the Apple Health to Obsidian guide walks through the import and the scripts that turn the JSON into the tables Obsidian indexes.

How often should I export?

As often as the question demands. Daily if you are actively investigating something. Weekly or monthly if the trend is the point. The export is cheap; the discipline of reviewing it on a cadence is the harder part.

Related guides

• Apple Health export to JSON, the schema reference for every destination on this page.
• Apple Health to Obsidian, the daily-note setup in detail.
• Analyze Apple Health with AI, the smallest possible setup, no scripting.
• Apple Health XML to CSV, the spreadsheet-first path.
• Build a personal Apple Health dashboard, the three-architecture decision tree.
• Apple Health for self-hosters, the homelab end-to-end pipeline.
• Bring a clean trend to your doctor, when the destination is a clinician.
• Health data portability checklist, the cross-cutting checklist for any destination.
• Export HRV and sleep data from Apple Health, the two metrics most quantified-self setups anchor on.
• Every metric HealthSave exports, the per-metric identifier catalog.
• Apple Health for cyclists, the meta-intent page for cycling workouts, FTP, power, and cadence.

Get HealthSave

Free to download, no account. Try the JSON export on a single 7-day window, paste it into the AI of your choice, and ask one real question. That is the cheapest possible start.

Download HealthSave on the App Store