Eight minutes tracing how perception begins
The problem
The subject is pre-verbal experience, so anything that reads as an interface breaks it. That rules out progress bars, navigation, settings, and the affordances that make software feel like software.
What we built
Raw WebGL2 with a downsample pyramid and frame-rate-independent timing, so the piece runs at the same pace on a phone and on a workstation with no rendering library in between.
A low-pass heartbeat with a tempo ramp drives the sequence through the Web Audio API. The visual stages follow the audio timeline.
Camera stages use on-device inference and IndexedDB. Nothing is uploaded and there is no analytics, because a piece about pre-verbal interiority cannot also be a data collection exercise.
What this demonstrates
Building against the metal when the medium demands it.
Get in touch
Tell us what you are trying to build and what is in the way. We will tell you early if we are not the right fit.