Your iPhone has been quietly collecting data about your body for years. Sleep stages, resting heart rate, heart rate variability, steps, oxygen saturation — a rolling archive of how your biology is performing, day after day. It's all sitting in Apple Health right now. And if you're like most people, you almost never look at it.
Not because you don't care. Because the data is incomprehensible. A number that says your HRV was 42ms yesterday means nothing without context. Is that good? Bad? Better or worse than last month? Should you be worried? Should you push hard in training today?
This is exactly the gap that AI health advisors are designed to close — and why apps like Metrya exist.
Why Raw Health Data Doesn't Help You
The Apple Health app is a data warehouse, not a health advisor. It stores everything faithfully and displays charts — but it doesn't interpret. It won't tell you that your resting heart rate has been trending up for the past two weeks, that this correlates with the three consecutive late nights you logged, or that your current HRV is 18% below your 30-day baseline.
That kind of pattern recognition used to require either a personal physician who knew your baseline, a sports science consultant, or hours of self-analysis with spreadsheets. Most people have access to none of those things.
AI changes the equation. A language model that has access to your full HealthKit data can hold months of readings in context simultaneously — and answer questions about it the way a knowledgeable friend would.
What You Can Actually Ask an AI Health Advisor
The most powerful shift is moving from passive chart-reading to active questioning. Here are the kinds of queries that become instantly answerable when your health data is connected to an AI:
- →"Why am I so tired this week?" — The AI can cross-reference sleep duration, deep sleep percentage, resting HR, and HRV to give you a grounded answer rather than a guess.
- →"Is my HRV trending in the right direction?" — Not just today's number, but the 30-day arc and what it might indicate about your recovery and stress load.
- →"Should I do an intense workout today?" — A recovery score computed from your own baseline answers this far more reliably than a generic app recommendation.
- →"What does my sleep look like compared to last month?" — Trend analysis that contextualizes today against your personal history, not population averages.
- →"My doctor mentioned my resting HR — what should I know?" — AI can translate clinical numbers into plain language and help you ask better questions at your next appointment.
Important distinction: An AI health advisor is a data analysis tool, not a diagnostic tool. It helps you understand patterns in your own numbers. It is not a substitute for professional medical advice. When something looks concerning, the right response is always to consult a doctor — but going in with clear data and specific questions makes that conversation much more productive.
How HealthKit AI Actually Works
Under the hood, the process is straightforward. An app like Metrya reads your authorized HealthKit data — the types you explicitly permit — then packages the relevant subset into a structured prompt for a large language model. The model reasons over your data and returns a plain-English response.
The key technical distinction that matters for privacy is where the AI computation happens and who holds your data. Some apps funnel your health data through their own servers, store it, and then send it to an AI. Others — using a Bring Your Own Key (BYOK) model — send your data directly from your device to the AI provider you choose (Anthropic's Claude, OpenAI's GPT-4, or Google's Gemini), with no intermediate server in the loop.
Metrya uses the BYOK approach. Your health data goes from your iPhone directly to the AI provider whose key you've entered. The app doesn't operate servers. Nothing is stored outside your device.
The Metrics That Matter Most
Not all HealthKit data is equally useful to an AI advisor. The highest-signal metrics for personal health understanding are:
Heart Rate Variability (HRV)
HRV — the variation in time between successive heartbeats — is one of the best proxies for your autonomic nervous system status and recovery state. A high HRV generally indicates good recovery; a suppressed HRV often signals stress, illness, or overtraining. But "high" and "low" are meaningless without your personal baseline. This is where AI earns its place: it can track your rolling average and flag meaningful deviations.
Resting Heart Rate
Resting HR trends are slow-moving but highly informative. A resting HR that creeps up 5–8 bpm over two weeks is one of the earliest signals of cumulative fatigue, illness onset, or overreaching in training. AI can spot this trajectory before you consciously notice it.
Sleep Architecture
Total sleep time matters, but sleep stage distribution matters more. AI can help you understand not just whether you slept enough, but whether your deep sleep and REM ratios look healthy — and how they're trending over time.
Recovery Score
A composite metric that weights HRV, resting HR, sleep duration, and sleep quality against your personal 30-day baseline. This single number gives you a daily readiness signal without needing to manually reconcile four separate charts.
Getting Started: Three Steps
If you want to start getting real answers from your Apple Health data today:
Frequently Asked Questions
Start understanding your health data
Metrya is free to download. Connect Apple Health, add your AI key, and start asking questions about your own data — on your device, with your rules.
Download Metrya — Free