Judge, Tag & Classify Items in Bulk (text or images)

Ask calibrated questions about many items at once — text, images, or both — and get, for every item, a yes/no probability, a pick from your options, or a score on your scale, each with a confidence. Use it to tag or flag photos (product category, price band, visible damage), screen or classify posts and messages against criteria, code survey or interview answers against a fixed codebook, score leads or options against a rubric, route support tickets, or judge several photos together as one item. It only answers the questions and options you give it: it doesn't find or collect the items itself, discover themes you haven't listed, or write explanations. Useful as the filtering step after a search or scrape: keep the real complaints, qualified leads or relevant posts out of hundreds of results, or pick the strongest of many drafts or options. Up to 500 items, 20 questions and 1,000 answers per call. Text questions run on Jev, TypeSafe's calibrated decision model. Cancelling a running job returns the answers that already finished.