Period trackers may not need machine learning, just basic statistics on your own data

A developer argues that period tracking apps marketed as AI-powered are really doing simple arithmetic, such as averaging recent cycle lengths. They contend the real concern is that many such apps require accounts and upload sensitive cycle histories to company servers. The developer describes Mori, a free Android tracker that stores everything on the phone and predicts using the median of up to six recent cycles. It shows a date range with a confidence label, and reports 'not enough data' when history is short or irregular. The author notes the deterministic design makes results repeatable and testable, and that the app works offline.