AI Insights Disclosure

Effective Date: May 19, 2026 · Last Updated: May 19, 2026

This page summarizes how CalmBP's AI Insights Suite works, what data we send to the AI provider, what we never send, and how to control it. The authoritative legal text lives in our Privacy Policy (section 13) and Terms of Service (section 13). This page is a plain-language summary written for the people using the app.

What AI Insights does

The AI Insights Suite is a CalmBP Plus feature. It gives you three things:

Daily Briefing. Once per day, a short text summary of your last 7 days of readings and lifestyle patterns.
Insights Chat. A few questions per day about your blood pressure, answered from your own data plus general lifestyle guidance. The chat does not diagnose conditions or give advice about your medications, and it will point you to your clinician for anything medical.
Weekly Recap. Once per week, a longer summary of how your week trended.

AI Insights is off by default. You turn it on through an in-app consent screen that names the AI provider and describes the data flow. You can turn it off again at any time in Settings → AI Features.

What gets sent to the AI

Every AI request sends the following de-identified data:

Your blood pressure values and the date/time of each reading
Context tag categories you chose for readings (for example, "caffeine," "stressed," "post-walk")
Sleep, step, and walking-workout totals from Apple Health
That you take medications and whether you took them, not the medication names

In addition, when you type your own question in Insights Chat, the words you type are sent to the AI provider so it can answer your question from the de-identified data above. Your typed question is not stored and is not added to the de-identified data above. Please avoid typing personal details you would not want to share.

What does not get sent

Your name or email
Your account identifier or device identifier
Your IP address
Your location
Your medication names
Free-text notes you've added to readings
Voice-entry audio (transcription happens entirely on your device)

Before sending, we remove direct identifiers as a safeguard. CalmBP does not collect most of these in the first place.

Who the AI provider is

We send AI requests to Google's Gemini API on Google's paid tier. Per Google's published terms, paid-tier submissions are not used to train Gemini models. Google acts as our service provider under written terms and may not use the data for its own purposes.

What we log on our side

We log technical metadata about each request, such as timing and whether a response was returned. We do not store your prompts. To avoid making the same AI request twice, a response may be cached on our servers for a short time and then deleted. Chat conversations are not saved as history.

Safety filters

Every AI response runs through a safety filter before you see it and carries our standard medical disclaimer.

Not medical advice

AI outputs are informational only. They are not medical advice, diagnosis, or treatment. Do not start, stop, or change any prescribed medication based on AI output. Do not delay seeking care from a licensed clinician based on AI output. In an emergency, call 911.

How to turn it off

Open the app, go to Settings → AI Features, and toggle off. All further AI calls stop immediately. We retain the metadata logs described above per the retention schedule in our Privacy Policy (section 7).

Usage caps

Per-feature daily and weekly caps apply to keep costs predictable, and may change.

Questions

Email [email protected].