Guide
The biohacker's manual for AI health chat
Apple Health already has the raw material: sleep stages, HRV, workouts, symptoms, nutrition, meds, and heart metrics. AI Health Export turns that pile of charts into a conversation you can actually use.
Why charts are not enough
Apple Health is excellent at collecting data. It is less useful when you want to know why your recovery is slipping, whether late workouts hurt sleep, or what changed before a symptom flare. Those are not chart questions. They are investigation questions.
AI health chat works best when it becomes your personal health data analyst. You ask a specific question, the AI reads the relevant Apple Health metrics, and you keep narrowing the answer until you have a useful next step.
Ask plainly
Use normal questions instead of digging through every chart.
Find patterns
Connect sleep, training, HRV, vitals, symptoms, and food.
Repeat the loop
Follow up with better questions as the picture gets clearer.
What AI Health Export adds
AI Health Export connects to HealthKit on your iPhone, supports 172 Apple Health metrics, and gives you two useful paths: built-in Pro chat powered by Google's Gemini API, or clean CSV export for ChatGPT, Claude, Gemini, spreadsheets, Python, R, or a doctor-ready summary.
Built for repeat health analysis, not one-off files
Pro chat includes persistent chat memory that can remember health context like medications, conditions, goals, routines, and patterns across sessions. That matters because the best health questions are rarely isolated. They build on the last answer.
The operating loop
Treat AI health chat like a small weekly review. The goal is not to get a dramatic prediction. The goal is to make better decisions with the health data your iPhone is already collecting.
Start with the outcome
Ask for a decision, not a vibe. Better questions sound like: should I lower training volume this week, what changed before my HRV dipped, or what should I bring to my doctor?
Give the AI the right health window
A seven-day export is good for quick checks. A 30 to 90 day window is better for recovery, training, sleep, and symptom patterns. Longer windows help when you are looking for seasonality.
Ask for evidence, then follow up
Have the chat name the exact metrics, dates, and pattern it used. Then ask the next question. AI health chat becomes useful when it turns a chart into a back-and-forth investigation.
Prompt playbooks worth saving
Start with these, then make the follow-up question more specific. The winning move is usually asking the AI to cite the metrics behind its answer, then asking what would confirm or disprove the pattern.
Sleep and HRV
"Compare my sleep duration, deep sleep, REM sleep, resting heart rate, and HRV over the last 30 days. What pattern best explains the nights where recovery was strongest?"
Training load
"Look at workouts, active energy, exercise minutes, resting heart rate, and HRV. Am I building fitness, maintaining, or drifting into under-recovery?"
Daily behavior
"Which weekday patterns show up across steps, workouts, caffeine, sleep, and heart metrics? Give me the behaviors that appear most connected to better next-day recovery."
Doctor summary
"Summarize the last 90 days of my Apple Health data in a clinician-friendly format. Include notable trends, outliers, symptoms, medications, and questions to ask at my appointment."
What to ask across 172 metrics
The power is in cross-metric questions. AI Health Export can work across activity, heart, sleep, respiratory, nutrition, mobility, reproductive health, mindfulness, 36 clinical symptoms with severity, and iOS 18+ event types. See the full metric list.
Heart and recovery
Resting heart rate, HRV, VO2 max, blood pressure, blood oxygen, respiratory rate, ECG, and workout recovery context.
Sleep and routine
Total sleep, time in bed, core sleep, deep sleep, REM, sleep stage timing, bedtime drift, and next-day recovery effects.
Training and output
Steps, active energy, exercise minutes, stand hours, distance, running power, cycling, swimming, and detailed workout types.
Body, symptoms, and notes
Weight, BMI, body fat, nutrition, hydration, caffeine, glucose, insulin, medications, symptoms, and appointment-ready trend summaries.
In-app chat or CSV export?
Both are useful. Choose the path based on how much control you want and how repeatable the workflow needs to be.
Built-in AI Chat (Pro)
Best for fast, repeatable analysis without leaving the app. Your health metrics are sent to Google's Gemini API only when you actively chat, and Google does not use API data to train models.
- No file cleanup or upload flow
- Built-in prompts and follow-up chat
- Chat history stays on your device
Export CSV to any AI tool
Best when you want portability, charts, deeper custom reports, or a model outside the app. Export clean CSVs, then upload them to ChatGPT, Claude, Gemini, spreadsheets, Python, or R.
- Use the model or tool you prefer
- Keep your analysis portable
- Decide exactly what to share
Privacy boundaries to understand
Health data deserves plain language. AI Health Export does not run a hidden health-data backend, and it does not automatically upload your Apple Health export anywhere.
Your data, your pipeline
- Free and Premium exports are processed locally on your iPhone.
- Pro chat sends data to Gemini only when you actively send a chat message.
- Exporting a CSV to ChatGPT, Claude, Gemini, or another tool is your explicit choice.
- Face ID protection helps keep your health data protected on device.
Start chatting with your health data
Free to start. Analyze 7 days of data across 20 metrics. No account required.