Cadenza

Input
Cadenza asks for six small, sensory things — the last thing that made you laugh, what this morning smelled like — and listens to the way you answer rather than the answer itself. It measures the cadence of a voice: its pace, its pitch, how steadily it holds a note. All of it is analysed live on the device — nothing is recorded, transcribed, or sent anywhere. If you'd rather not speak, you can type, and it reads the rhythm of your keystrokes instead.

Transformation
From that reading Cadenza weaves a resonance bloom — concentric, wandering loops laid down on paper, where loudness sets their weight, unsteadiness sets their wander, and register drifts the palette from warm to cool. The plate is then marked the way one marks a passage of music: a tempo, a dynamic, an articulation — Andante · mezzo-forte · legato. The same reading always yields the same plate; no two readings fall alike.

Note
A voice is the most personal thing the browser will accept; a musical marking is the most impersonal record of one. Cadenza performs the translation between them and keeps neither end — not the answers, not the voice, only the shape of the voice while it answered. What remains is a portrait made from the parts of a person that can be measured, and marked in a language that offers no verdict about what it heard.