Skip to main content
Typed decision models judge text and answer with something your app can act on: a category, a score, or a probability between 0 and 1. They do not write sentences, so keep replies, summaries, and free-form extraction on a chat model. The typed decision model is Jev 1.13, from TypeSafe. You give the app some text, such as a ticket, a comment, or a list of candidates, and the questions you want answered. Ask Lovable to:
  • Route support tickets to a team or queue, with a catch-all when nothing fits.
  • Flag comments against several independent policy rules.
  • Rank search results, products, or candidates by relevance.
  • Score quality on separate dimensions, then combine the scores in your app.
  • Check a proposed answer against source text you already retrieved.
Ask for any of these in the project chat, and Lovable writes the feature and the backend call behind it. See Add an AI feature to your app for the whole flow, including publishing. Jev evaluates text only. For images, audio, or video, transcribe or describe the content first. You can also use a chat or embedding model that accepts that media. Jev is trained primarily on English, so test other languages before you rely on them. Jev 1.13 is not available in workspaces that have EU inference enabled.