Tap words to build a message, then Speak. Nothing else is asked of you. Every tap's timing and precision is measured; when your precision drops below your own baseline, targets grow, spacing widens, and a hold-to-select ring appears.

Gait from the phone's motion sensor

Start, put the phone in a front pocket or hold it still against your body, and walk for a minute. Step timing, cadence and stride regularity are computed on the phone as you go.

A baseline is set from your first 30 steps in this session. Later windows are compared with that baseline only — never with a population norm.

Posture and arm swing from the camera

Prop the phone up so your whole body is in frame and walk toward and away from it. The pose skeleton is computed on the phone; no video is recorded or sent.

Toe-walking and alignment are approximate from a single camera; the native app will use the same primitives with calibrated distance.

Drawing, doodling, handwriting

Draw anything with a finger. Stroke smoothness, speed steadiness and in-stroke tremor are computed from ordinary doodling — the fine-motor record an OT never gets between visits.

Talking

Read anything aloud or just talk. Speech rate, pause ratio and pitch variability are computed live on the phone. Only these numbers are kept; the microphone audio is never stored.

For people who do not speak, the Speak board carries the SLP measures instead: selection time, message length and vocabulary breadth.

Readiness — The Ready Mindset

Movement tells us what the body is doing. Readiness tells us what the person or caregiver can carry right now. The Ready Mindset (MQ) profile shapes how PRAXIS phrases every suggestion — pace, tone, what to lean on — so the plan fits the person, not just the muscles.

Your own baseline — one record, three lenses

PT reads the gait columns, OT the stroke and tap columns, SLP the speech and board columns — from the same saved sessions. Your first three saved sessions become your baseline. Every later session is judged against that — steady, shifting, or changed — and against the sound level it was recorded in. No population norm is used anywhere.

Explain this in plain language

A language model reads the latest saved session, your baseline and your readiness profile, and writes it for the audience you pick. It may only use the numbers; it never names a condition or prescribes.

Precision by environment

Sessions on this phone

One nervous system, one arc

Infant movement, childhood praxis, adult catatonia risk and late-life parkinsonism are studied as four separate subjects. PRAXIS measures them as one arc. Tap a point.

What the arc lets us ask that nobody can today

Does the two-year-old whose reach paths straightened slowly become the fourteen-year-old whose transition latency lengthens under load, and the forty-year-old whose stride regularity drifts a decade before a parkinsonism diagnosis? No dataset on earth can answer that, because no one has measured the same primitives across the whole span. PRAXIS is built so that it can.

PRAXIS Motor Phenotype Ontology v0.1

An open, versioned vocabulary for computable phenotypes across the lifespan, read through three lenses — PTOTSLP — from one stream. Each primitive names its everyday source — the activity it is measured from without asking. Marked primitives are live in this demo.