How Well Rhythm Works — Basis & References

1How cycle prediction works

Predictions are computed statistically from your own past records (the average and variability of your cycle lengths) — not guessed by an AI or inferred from other people’s data. The more you record, the more the prediction fits you.

Public references: basics such as typical menstrual cycle ranges follow definitions and educational material from the Japan Society of Obstetrics and Gynecology and the MHLW women’s health resources.

Basal temperature & pregnancy: the basal temperature chart follows the common idea of a biphasic (low- then high-temperature) pattern. In pregnancy mode, gestational week and due date are calculated from the last period or due date you enter — this is not a diagnosis.

2About barometric pressure and how you feel

That weather changes (especially falling pressure) can act as a trigger for headaches is mentioned as one of the trigger factors in clinical guidelines such as those of the the Japanese Headache Society (“Clinical Practice Guideline for Headache 2021”), and remains an active research area. However, individual differences are large — it is not a rule that applies to everyone.

Honest disclosure: the thresholds behind indicators like "rapid pressure drop" are Well Rhythm’s own design values, informed by public research and general meteorology — they are not medical diagnostic criteria. Weather data comes from Apple Weather. Pressure is taken from weather data (mean sea-level pressure), not from your device’s sensor.

About “pressure sensitivity”: the sensitivity you can set is a simple guide based on your own choice and records — not AI learning (it is not a learning feature at this time).

3Your data is the primary evidence

At the center of Well Rhythm is your own record, not general theory. The correlation view computes the relationship between your symptoms, pressure, and cycle entirely on your device. Whether low-pressure days really mean more headaches for you — your data holds the answer.

Cycle predictions also come with a self-check: you can see how far past predictions deviated from what actually happened. We hide neither the hits nor the misses.

How correlation is measured: the analysis counts, on your device, how often two records overlap — for example “days pressure dropped” and “days you had a headache” (a relative frequency). That’s why it only becomes meaningful once you have enough records (about a week or more as a rough guide). And correlation is a hint that things “may be related” — not proof of cause. Even if low-pressure days seem to bring more headaches, we don’t claim that as the cause.

About the moon phase: we include the lunar phase as one lens for correlation, but make no medical claim of causation. It’s an optional way for you to check whether there seems to be a relationship in your own data.

4About the daily message

The daily message is a self-care template chosen from your records and the pressure forecast. It is not medical advice and promises no therapeutic, diagnostic, or preventive effect.

How we use AI, disclosed: we may use AI for things like gently rephrasing text, but the default is templates. When AI is used, it runs on your device (offline) and sends no health data outside. It gives no medical advice, and its output passes through safety rules designed not to encourage diagnosis or treatment before it is shown.

5What Well Rhythm does not do

• No displays intended for diagnosis, treatment, or prevention (it is not a medical device).
• No claims that guarantee effects.
• No health data sent to our servers (encrypted on your device). See the Privacy Policy.

If you feel unwell, please consult a medical professional. This page is reviewed and updated whenever features change.