How it works
From a daily check-in to a forecast you can question.
HerPattern is a research prototype that combines your own logged history with population-level patterns to forecast symptom burden. It always shows its uncertainty and never diagnoses or recommends treatment.
1. Daily check-in
Log symptoms with quick chips, an optional note, or a voice note, usually under 30 seconds. Voice extractions are always shown back to you for confirmation before they're saved.
2. Personalization
As your history accumulates, HerPattern's models are partially pooled across a population baseline and your own pattern, so forecasts adapt to your individual cycle and symptom history rather than treating everyone the same.
3. Forecast for the next 3 days
You see the probability of a low, moderate, or high burden day for each of the next three days, an 80% plausible range, and a plain-language confidence label. Never a single confident number.
4. Reports
Generate an appointment-ready summary of your recent pattern, symptoms, and forecast history to bring to a clinical visit or keep for your own records.
Forecast readiness
Forecasts need enough recent history to be meaningful. Early on, or after a gap in logging, HerPattern will tell you a forecast isn't ready yet rather than guessing with false confidence. As your check-in streak grows, confidence typically improves.
Missing data
Missed days are common and expected. The model accounts for gaps explicitly. It widens the uncertainty range rather than filling in guesses, and the app flags when a gap is likely affecting forecast quality.
Forecasts can be wrong
HerPattern's forecasts are probabilistic estimates, not guarantees. They are not a diagnosis and should never replace clinical judgment. Use them as one input among many when planning your day or preparing for an appointment. See our methodology and medical disclaimer.