Tell me what you need the camera to understand. I’ll ask questions when the task is ambiguous, then build and revise the active model as we test it.
01 / COLLABORATE
Build through conversation.
Describe the outcome, constraints, and what you see in the results. The agent can build and revise the active visual model as you test it.
EARLY RESEARCH ACCESS
Sign in to test the compiler.
Build classification, detection, and segmentation tasks through conversation, then test them with a camera or image.
Open registration. Usage is rate limited while the models are under active development.02 / TEST
Live model output
Loading checkpoint