Camera processing happens on your machine — and only while a session is active.
A short, plain-language summary of what the camera does, what data exists, and how to remove it.
The webcam stream is read frame-by-frame inside your browser to extract a small set of numerical signals (attention, calm, engagement, discomfort).
Models run on your device. The video stream is never sent to our servers and is released the moment the session ends.
Only your final result — predicted colour, design ranking, confidence — and only if you choose to save it locally.
No raw video. No frames. No face embeddings. No identity. No location. No telemetry from the session.
All saved reports live in your browser. A single button in the Reports page clears everything immediately and irreversibly.
Colour sessions and design reviews start only after camera permission is granted. The stream stays local and is released when the session ends.
THE MAZE produces estimates of preference based on visible response. It is not a medical or psychological assessment. It does not diagnose conditions, infer identity, or provide clinical guidance. Use it as a thinking aid, not as a verdict.