This Hugging Face Space pairs two models. YOLO-World finds boxes around the things you name in a photo, and Jev answers one question about each box: should it be kept? Kept boxes and skipped boxes are drawn in two different colours.
You choose the classes to look for, can add a scene hint, and set two sliders: how sure Jev must be to keep a box, and how low YOLO's own detection floor goes before Jev sees the boxes. A table lists every detection with both scores.
It runs in the browser, so you open the Space and try it. To run your own copy you add a TypeSafe API key as a Space secret, or paste one into the page for a single session, and the page says it is not stored.
What you get
- A running app on Hugging Face you can use in the browser
- The Python source, readable and copyable from the Space's Files tab






