Your cycle, not the average
Once there is enough personal history the app does not fall back to a generic 28-day cycle. Someone recording 34, 35, 36, 34, 35 and being predicted 28 is treated as a product defect, not a tuning issue.
Private health tracking
Privacy: Period Tracker
Your cycle belongs to you.
A private Android period tracker that learns your own cycle rather than the average person's, and keeps every date on your device.
Pre-release_> Android, Kotlin, Jetpack Compose, Room, offline-first> Privacy-first product direction> On track
What it is
Privacy: Period Tracker answers one question well: when is the next period likely to start? It records confirmed dates, learns the individual pattern from them, and produces a most-likely date with a window and a confidence label, never a bare date presented as fact. From there it estimates ovulation and the fertile window, and asks discreetly whether the period started, using both answers to sharpen the next forecast.
It is deliberately not a community, a pregnancy mode, a chatbot, an article feed, or a symptom encyclopedia. There is no account and no server: cycle history stays in a local database on the device.
Product loop
track -> learn -> predict -> remind -> learn again
Screenshots
The product surface.
What makes it different
Built with a clear Understanding.
An honest window
Every forecast carries a range and a confidence label instead of a single date presented as certainty.
Both answers teach it
The discreet check-in feeds yes and not yet back into the model, so a period that has not arrived is information rather than a missed notification.
No data sale, held by the build
The promise never to sell cycle history, fertility information or predictions is structural: a module boundary check in the Gradle build grants the future advertising module no project dependency at all.
Platform notes
The practical shape of the build.
A quick look at how the product is built, what is available now, and what stays intentionally limited.
Kotlin and Jetpack Compose on Material 3, Room for the cycle database, DataStore for preferences, and pure-Kotlin domain modules that cannot see the Android SDK.
No account and no server. Cycle history, predictions and fertility estimates stay on the device, and none of them may be written to a log, an analytics event, or an ad request.
Ovulation and fertile-window estimates are predictions from recorded history. They are not a contraceptive method and must not be relied on as one.