Vercel put Jev behind a plain HTTP endpoint that returns probabilities instead of text.

AI Gateway now exposes TypeSafe's decision model three ways: point an existing TypeSafe client at /typesafe, POST to /v1/evaluate, or call experimental_evaluate in the AI SDK. You send state plus a questions object and get typed answers with probabilities attached, boolean returning 0 to 1, choice selecting from named options, score rating against a defined scale, which removes the parse-and-validate wrapper around most LLM classification calls today. All three paths bill through AI Gateway, so decision-model spend appears in the same usage views as token spend. (Vercel)