← back to legend
Ants: routing and our own terms (from Nov 2026)
VLM
Vision-language model
An AI that can look at an image and describe or judge it. Useful as a reviewer of a render, not as a router.
How it looks
Drawing: an AI tool, not part of a board
Illustration of the idea: a picture of a board goes in, a written judgement comes out. The words are an invented example, not a real review.
How it works
Where it fits
A possible reviewer of the 3D render made at the verify stage, next to KiCad's DRC and a human.
Contrast with a learned (RL) router, which is also AI but is trained to place copper itself.