Aerendir · Cycle Wellness

Aervulate

Help us build a cycle wellness app using the motion sensors in the phone you already have. Five seconds, three times a day, could help us learn what the signal can tell us.

A simple recording today. Better cycle insights are the goal.

Phone-based data collection

Record phone motion data without additional wearable devices or test strips.

Brief recordings

Hold your phone still for five seconds, morning, midday and evening.

Just your phone

Compatible phones can record small hand movements with built-in motion sensors.

Restricted data access

We do not ask your name. Your email is kept separately from recordings.

The science

Your hand moves in tiny ways you may not notice. We are exploring whether phone measurements of those movements show useful patterns across the menstrual cycle.

Illustrative signals only. These are not measured results or evidence of cycle detection.

With three brief recordings a day across roughly two to three cycles, we can evaluate whether the signal is useful for a future cycle wellness app. The current version collects recordings. It does not yet identify ovulation or provide fertility guidance.

How to participate

Three recording windows each day.

Scan & enroll

A QR code opens Aervulate on your phone. Confirm your email and receive a participant number. Signup asks your age, when your last period ended, your height and weight, smoking status, and hormonal birth control use.

Hold still for five seconds

Morning, midday, evening. Hold your phone upright and still in one hand, in the air. Keep your arm and elbow unsupported, away from your body and furniture. The page guides each five-second capture.

Carry on for two to three cycles

Continue for approximately 80 days. Each saved recording helps us build a dataset across time.

Collection overview

2,500volunteers · 18–50 · any cycle, regular or not

3×daily captures · slots a · b · c

5 sper capture

~80days · 2–3 cycles

~60 Hzbrowser event rate varies by phone

Full applifetime access after release for participants who complete the collection requirements

We never ask your name.

We do not ask your name or precise location. Your recordings use a participant number. Your email is stored separately, used to verify sign-in, and excluded from recording exports.

Project access

The private prototype now covers joining, receiving a participant number, and recording a session.

✿ Join Aervulate Join with an email, receive a participant number, and record five seconds. Open on a phone. →

⚙ For the team — data model & roadmap

Table	Fields	Notes
`collection.user_data`	`user_id`, age, height, weight, bleeding end date, smoking, birth control, exercise days	Signup answers. Email and authentication stay in the restricted identity schema.
`collection.form_data`	`form_id`, `user_id`, recording and receipt times, slot, browser, acceleration source, bleeding answer, sample format	One row per recording, linked to its participant.
`collection.raw_data`	`form_id`, `sensor_type`, `event_index`, `t`, `handler_time_ms`, `reported_interval_ms`, `x`, `y`, `z`	Current format: browser event and receipt timestamps in milliseconds, gyro in degrees/s, acceleration including gravity in m/s². Timestamps describe browser delivery, not exact hardware measurement times. Older recordings use the units identified by their sample format.

PostgreSQL links users to recordings and recordings to sensor readings. One saved recording per participant, local date and daily slot.

done Email-verified join and returning sign-in, issued participant IDs, phone recognition, collection database, capture flow, and admin session browser.
done Hold-still pose on the recording page, Home Screen setup, and reminder sends.
next Review IMU capture and collection criteria with Martin.
done Five-second captures, numeric participant and recording IDs, and versioned raw sensor storage.
later Confirm hosted email verification, schedule reminder checks, configure the participant host, and approve participation wording.
later The 5–7 slide volunteer deck, then outreach; training begins at 2,500 enrolled.

Reference: Martin's email of 2026-09-03, verbatim in the repo at docs/AERVULATE_REFERENCE.md. That email is the source of truth. Martin leads the feature search after collection.