VEXA Cosmic Crew: a live smoke simulation warping daily NASA feeds.
Today's drop is loading...
Source:
How this piece was madeThe world builds in public. Check back daily.
Experimenting began June 23, 2026.
Custom-built code pulls the NASA source image and runs it through VEXA's smoke training data: the same hand-shaped visual language that defines the portrait series. No generative AI. The output is what the data feels like, translated into smoke and form.
Nothing here is planned in advance. The palette that emerges from 30 drops will be different from the one at 100. This is a living record of VEXA's color intelligence evolving in real time, shaped entirely by what the universe is saying on any given morning. Come back. Watch it grow.
The body is the photograph's own light. Each piece reads the brightness structure of that day's NASA image and lets the smoke find its figure there. The shape is never drawn directly; only felt, bent into the color the signal becomes.
About the data
NASA imagery is sourced via NASA's open APIs (NASA Image Library · APOD · EPIC). Only public-domain images are used: APOD sometimes features photographs owned by individual astrophotographers or observatories, flagged by a copyright field on the image. Those are refused at the moment of fetching, so a copyrighted photograph cannot enter the pipeline at all. Copyrighted images are refused at the source; only public-domain NASA imagery can ever enter a piece.
Not affiliated with or endorsed by NASA.
Hover to magnify · Esc or click outside to close
