A short face scan and a one-minute voice task, read by an experimental engine that writes a reflection in plain words. No account, no name, no questionnaire.
Everything about your face is analysed on your device. Your voice recording goes to Numa's server for transcription and is deleted straight after.
Numa is an engine that listens to how you look and sound right now, and tells you what it noticed, in language a friend would use.
Modern models do the reading: on-device face tracking, speech recognition, and a language model that writes, never a questionnaire.
The subject is you, not a category. Every reflection is written from your own words and signals, and rated by you afterwards.
Face and voice together: head movement, expression, blink, pitch, pace, pauses, energy, and what you actually said.
Signals become five vectors and a narrative: a signature, your strengths, one blind spot, and how today differs from last time.
Every step waits for you. Nothing is recorded unless you actually did it.
Look ahead, turn slightly left, then right, come back to centre. Face tracking runs on your device and the scan only counts if it saw you do it. Frames never leave the phone.
Five short lines to repeat, then one question on its own screen: What will you do tomorrow? Forty-five seconds, in your own words.
Your recording is transcribed and reduced to some seventy acoustic parameters, then deleted. Face summaries, voice summaries and your words are fused into five vectors and a written reflection.
Come back tomorrow, next week, next month. Numa compares each session with your last one, so the story is about change, not a verdict.
Each one has a trait reading, by design, and a state reading, right now. Scores stay in a moderate band on purpose: one session is never an extreme verdict.
How easily things move through you today: engagement, rhythm, momentum.
Steadiness of attention: how settled you look and how cleanly you speak.
Your reach toward the world: expressive range, warmth, the will to be met.
Energy and recovery: what is in the tank, and how fast it refills.
How much of the inside shows on the outside, in face and in voice.