← Back to VitalBrief
Apple Health + AI

AI-powered insights from your Apple Health data.

Apple Health stores years of data most apps barely touch. VitalBrief reads it, runs it through an AI, and writes the kind of summary a thoughtful coach would write, weekly and monthly.

What an AI sees that a dashboard does not

A dashboard shows numbers. An AI report shows the relationships between them. That is the difference, and it is more useful than it sounds.

Examples of what shows up in real reports:

  • HRV consistently lower on the days following long Zone 2 sessions, and recovering when a rest day is added.
  • Resting heart rate trending up across the week alongside two short nights of sleep.
  • A drop in workout volume that matches a stretch of higher than usual stand hours, suggesting more activity overall but less structured training.
  • A peak training week where every metric agrees, and the Wellness Score reflects that.

None of these are interesting on their own. All of them are interesting when seen together, which is the point.

What gets sent and what does not

The app does not add personal identifiers to report requests. There is no name, no email, no device ID, no account ID added by the app. What goes through the proxy is a structured summary of the metrics the report needs: a few weekly aggregates, daily values for HRV and resting heart rate, workout summaries.

The AI service receives only that summary. It is configured for zero data retention where supported. The proxy processes limited technical metadata needed to route the request, debug errors, and prevent abuse.

Your full Apple Health history, reports, and profile are stored on your device.

Wellness, deliberately not medical

VitalBrief is a wellness app. The reports use words like trends, patterns, observations, and insights. They never use diagnostic language. If you want a medical reading of your data, that is a conversation for a qualified healthcare professional, with their own tools and your raw numbers.

What the AI is good at is being a patient reader of long-running data. That is most of what we want from this kind of feature, and very little of what most fitness apps actually do.

What kinds of insights you get

The reports adapt to what the app knows about you. If your profile says you train mostly running and cycling, the language and benchmarks are different than if you mostly do strength training. Sport-aware context is built into how the prompt is constructed, not bolted on after.

The cadence is two reports plus an onboarding one:

  • Welcome report: looks at the past 12 months in monthly slices, generated once after onboarding.
  • Weekly report: 200 to 350 words, written every week.
  • Monthly report: 400 to 600 words with a month-over-month comparison table.

Frequently asked questions

What AI does VitalBrief use?

The app is provider-agnostic. Requests go through a Cloudflare Worker proxy that talks to an external AI service. Specific providers can change without affecting how the app works. The user-facing language is "an external AI service" and that is honest.

Will the AI diagnose problems?

No. The reports stay strictly in wellness language. Trends, patterns, insights, observations. Never diagnoses, detection, or risk.

Where are reports stored?

Reports and profile are stored on your device using Apple frameworks. When a report is generated, the AI service receives a structured summary without identifiers added by the app and is configured for zero data retention where supported.

What if my Apple Health data is sparse?

The app evaluates eligibility before every generation. With partial data, you get a shorter, lower-confidence report that explicitly notes what is missing. With too little, the app skips the AI call and shows guidance instead.

Does optional "About You" text get sent?

Yes, only the text you typed yourself, sanitized before being included in the prompt. It is capped at 500 characters. If you do not enter anything, nothing extra is sent.

Read your Apple Health data, finally.

VitalBrief is launching on the App Store soon.

Coming to App Store