Privacy policy

Last updated

This policy is a draft: it has not been reviewed by a lawyer, and it will change before the app is released. It covers two separate things: this website, and the healthd iPhone app.

The website

This section describes the website as it is built today.

This site is hosted on Cloudflare, which serves its pages from Cloudflare’s network. Cloudflare handles every request, including your IP address and the details of the request, to deliver the site and to protect it, for example with its browser integrity check, under Cloudflare’s privacy policy. The site keeps no request logs of its own and sends none anywhere.

Cloudflare Web Analytics is turned on for this site: it may count page views without cookies, recording the page, the referring page, the country, the browser and device type and how fast the page loaded, without identifying you and without storing anything in your browser. Until you accept cookies in the banner, the pages set no cookies and load nothing from another server, except Cloudflare Web Analytics when it is active.

If you accept, two more tools load. Google Analytics 4, from Google, records the pages you visit, how long you stay, your approximate location, your device and browser and a pseudonymous identifier for your browser; it sets the cookies _ga and _ga_<id>, which last up to 2 years, and _gid, which lasts 24 hours, where Google sets it. That data goes to Google Ireland Ltd and Google LLC and may be transferred to the USA, under Google’s privacy policy. The Meta Pixel, from Meta, records the pages you visit and browser data linked to a pseudonymous identifier, which Meta may match to a Facebook or Instagram account; it sets the cookie _fbp, which lasts 90 days. That data goes to Meta Platforms Ireland Ltd and Meta Platforms, Inc. and may be transferred to the USA, under Meta’s privacy policy. We use them to see how visitors find the app and to measure our ads.

Your choice is kept in this browser’s local storage under the key healthd-consent, not in a cookie. To change it, use Cookie settings at the bottom of any page: it forgets your choice, deletes the _ga, _ga_<id>, _gid and _fbp cookies, reloads the page and asks again.

There are no forms, and the site asks you for nothing.

The app

What the app reads

The app reads only the Health data groups you turn on in the app. It asks Health for read access only, and it never writes to Health.

Each group you turn on lets the app read these types of Health data, and no others:

Where your data goes

The app sends what it reads to one place: the receiver address you enter. Its only network code is the delivery to that address and the local-network check described below; separately, the paid features use Apple’s App Store purchase system, which talks to Apple (see In-app purchases). The app as released contains no web address of its own, so there is no server of ours and no third party it could send your Health data to. A plain http:// address is accepted only on your own local network; any other address must use https://. Each request carries the samples from the groups you turned on, with the method and headers you set, including your credential. The app never reads the body of your receiver’s response, only its status code, and it does not follow redirects. It sends with a network session that keeps no cookies and no cached responses on disk. If you point the app at a service that someone else runs, what happens to the data there is up to that service.

What the developer receives

No Health data, ever: there is no server behind the app, and there is no account to create or sign in to. The app contains no analytics, advertising, tracking or crash-reporting code. It uses no third-party libraries, only Apple’s own frameworks. Its privacy manifest declares no tracking and no collected data types. The app has no iCloud capability and no iCloud code, so it stores nothing in iCloud itself. Separately from the app, iOS can share crash reports with app developers if you allow that in the iPhone’s settings, and TestFlight shares crash reports and feedback with the developer of an app you test; those are Apple’s features, under Apple’s settings.

In-app purchases

The paid features are bought through Apple’s App Store; Apple takes the payment, and the app never sees your card, name, email address or Apple Account. To know which features to unlock, the app asks the App Store, when it opens, which of its purchases are active on your Apple Account, and keeps the answer on your phone. Apple gives the developer sales and subscription reports. They carry no name, email address or contact details. They do show, for each purchase: the product, the price, the currency, the country or region of the App Store, the type of device, and, for subscriptions, a random subscriber number that Apple creates for each customer of each developer. Apple deletes that number if you have had no subscription for 180 days. What Apple itself does with your purchase is described in Apple’s privacy policy.

What stays on your phone

The app keeps your settings on the phone: the receiver’s address, method and headers, the groups you turned on, and your export and alert settings. It also keeps whether the paid features are unlocked, and whether they ever were, so automatic export can check without asking the App Store; whether it has already explained the Health and local-network permissions; and your silence alert settings. It keeps a delivery log of the last 100 runs: when each ran, what started it, how much was sent, the response status and the outcome. The log holds counts, not health data. Batches that have not been delivered yet stay on the phone, with the health data in them, until your receiver accepts them or for at most seven days. The same file records how far delivery has got, so the next export carries on from there. All of these are files inside the app, written through one place in the code. They are protected by iOS Data Protection and cannot be read until the phone has been unlocked once after it starts. Your credential is kept in the iOS Keychain, on this device only, and is not synced to iCloud Keychain. Once saved, it is shown masked and is never displayed again. The app’s only log lines are about background work and about a file it could not mark to be left out of backups, and they carry no credential and no health data.

Device backups

The file that holds batches not yet delivered is the only one the app marks to be left out of iCloud Backup and computer backups. The app sets that mark again every time it writes the file. If iOS refuses the mark, the app still saves the file and logs only the file’s name and an error code. Your settings, the groups you turned on, your purchase status and the delivery log are not marked, so they can be included in a backup of your iPhone made with iCloud Backup or a computer. If you restore your iPhone from a backup, the app no longer knows how far delivery had got, so it reads your Health data from the beginning again and sends your history once more. Your receiver is expected to drop the repeated samples by each sample’s identifier.

Local network access

If your receiver is on your own network, iOS asks whether the app may use the local network, because without that permission the app cannot reach it. The app explains why before iOS asks. To tell whether the permission is on, the app briefly announces a test service on your network and looks for it itself, for at most four seconds; if the first delivery of an export fails while iOS is still waiting for your answer to its prompt, the app looks again for up to 30 seconds and, if you allowed access, sends that delivery once more. Any connection made to that test service is closed at once. It runs this check when you test a receiver on your network, and once at the start of each export to one. A receiver on the internet needs no such permission, and the app does not ask for it then.

Notifications

Failure alerts are off until you turn them on, and only then does the app ask for permission to send notifications. They are local notifications created on the phone; the app uses no push service. An alert says how many deliveries in a row failed and what kind of failure it was, and shows no address, no status code and no health data. The silence alert is off until you turn it on, and only then does the app ask for permission to send notifications. It is a local notification scheduled on the phone, which fires if nothing has reached your receiver for the number of hours you choose, and it says only “Nothing delivered lately” and “Nothing has reached your receiver in N hours. Open healthd to see why.” Failure alerts are a paid feature; the silence alert is free.

Your choices, and deleting your data

You can turn any group off in the app, or revoke the app’s access in Settings → Privacy & Security → Health. Remove Credential, on the receiver screen, deletes the credential from the Keychain. Deleting the app deletes its files: your settings, the delivery log and any batches not yet delivered. iOS can keep a Keychain item after its app is deleted, so use Remove Credential first if you want the credential gone as well. If you install the app again, it deletes a credential it finds without settings. Data already delivered is on your receiver, which you run, so you delete it there; the app cannot delete it for you, and it never tells your receiver about deletions in Health.

Children

Nothing the app reads from Health reaches the developer, from anyone, children included.

Changes and contact

If this policy changes, the new version will be posted on this page with a new date. Questions about this policy go to support@healthd.app.