Splanc

LED mapping is a solved commercial problem, up to a point. Existing tools still ask a person to trace the fixture: click each point, drag a path, correct the mistakes. That loop is slow, and it is error-prone enough that the map often becomes another thing to maintain.

Splanc takes the human out of that loop. You walk around a lit sculpture, ceiling, tree, or room with a phone. Computer vision recovers every LED's position in three dimensions, then turns the cloud into strands and junctions. You do not trace the fixture. You capture it.

Mapping is only the first half. Once the geometry is known, effect authorship moves into an agentic harness. Instead of writing light as if it were still a strip of indices, you can verbally prompt the fixture to behave the way you want — and the system plays that behavior on the real object.

This is not an artwork. It is infrastructure for treating large-scale light as form, and for talking to that form directly.