Technical preview

Your health data.
Your AI.
Your control.

Bring your Apple Health data to your own cloud account, encrypted at rest — and share it, deliberately, with the AI assistants you choose.

Not a medical device · No availability claims until it ships

Three steps, no leaps of faith

1

Import from Apple Health

Your iPhone reads Apple Health on your schedule and queues day-level buckets in an encrypted durable store — nothing leaves the device until sync runs.

2

Encrypted cloud storage

Buckets sync to your own cloud account as ciphertext. Keys are managed server-side inside the security boundary; data older than 90 days rolls off.

3

Deliberate sharing

You grant each AI assistant access to explicit categories with an explicit lookback window. Revoke at any time; grants are checked on every request.

Nine metrics. Five categories.

The MVP scope is deliberately narrow: nine day-level metrics from Apple Health, nothing more. Sensitive categories are marked and shared only under an explicit grant.

Activity

Steps · Active energy · Distance

Heart

Sensitive

Heart rate · Resting HR · HRV (SDNN)

Sleep

Sensitive

Sleep duration

Workouts

Workout sessions

Body

Sensitive

Weight

Private by construction

Ciphertext at rest

AES-256-GCM with per-user data keys; the key-encryption master key lives only in server-side secrets. We call this server-managed encryption — not end-to-end — because we're precise about it.

90-day rolling window

Data older than 90 days is retired on schedule. Full JSON export and deletion are yours, behind explicit barriers — including account deletion.

Three separate consents

Reading Apple Health, storing in the cloud, and sharing with each AI app are independent permissions. Empty read access proves nothing; missing data is never treated as zero.

Encryption is server-managed, not end-to-end. We say so plainly, because a privacy product that blurs its own guarantees is no privacy product at all.

AI assistants on a short leash

Connect an AI client through the Model Context Protocol and decide, per app, what it may read: which categories, how far back, and whether it may ask at all.

  • Per-app grants with categories and lookback you control.
  • Every request is checked against the exact active grant — no standing access.
  • Revoke instantly; the next request is refused.

Example grant

assistant · “Coach”

Activity
HeartSensitive
SleepSensitive
Workouts
BodySensitive
Lookback30 days

On iPhone, when it's ready

The iOS app is in development. No pre-registration theater — just the App Store page, updated the day the link goes live.

Questions, answered straight

Is my health data end-to-end encrypted?

No — and we won't claim it is. Data is encrypted at rest with per-user keys, and keys never leave the server security boundary. That is server-managed encryption: strong, auditable, honestly named.

Which AI assistants can connect?

Any AI client that supports the Model Context Protocol once you grant it access. Each grant is explicit: categories, lookback window, and an on/off switch checked on every single request. External catalogs are off unless the owner enables them.

What exactly is shared with an AI app?

Only what the grant allows: the categories you picked, within the lookback you set. The app never sees your passphrase, your keys, or categories outside the grant.

Can I leave?

Yes. Export everything as JSON, delete the data, or delete the account outright — erase and delete are first-class flows, not dark patterns.

Is the iOS app on the App Store?

Not yet — it is in development. The App Store page will carry the link the moment it ships.

Is HealthBridge AI a medical device?

No. HealthBridge AI is not a medical device and does not provide medical advice. It moves and guards data you already own.

Start with your own data

Sign in to the web portal, connect your iPhone, and decide who gets to know what.

Sign in