Ask Jev Anything
Type any yes/no question — Jev returns a calibrated YES · NO · IT DEPENDS with the odds behind each choice. Not generated text: one typed decision in milliseconds.
What is "Ask Jev"?
Ask Jev is a tiny front-end over the Jev decision model. You type any yes/no question, and instead of writing you a paragraph like ChatGPT would, Jev returns a single typed verdict — YES, NO, or IT DEPENDS — with a calibrated probability behind each option. It's the same primitive developers call from code to route requests and gate actions, wrapped here as a toy so you can feel what a "decision model" does.
So this is not a chatbot and it doesn't generate text. It reads your question exactly as written and returns odds. Ask "Is pineapple good on pizza?" and you'll get something like IT DEPENDS 76% · YES 23% · NO 1% — a distribution, not a lecture. That's the whole point of Jev: a machine-readable answer your software could branch on, delivered in a few hundred milliseconds.
Why the answer is often "it depends"
Jev is calibrated, which means it tries to report honest uncertainty rather than sounding confident. Genuinely open questions land on "depends" with the probability mass spread out; clear-cut ones snap to a confident YES or NO. That spread is the interesting part — it's Jev telling you how sure it is, not just what it picked.
Want to build this?
Every result here is a real API call. The same endpoint powers routing, classification, scoring and safety gates in production apps — ask a choice, a score, or a calibrated yes/no (noul) and get a typed answer back. Grab a jv_live_ key and you can wire this into your own product in about 30 lines; the TypeScript quickstart has the code.
FAQ
Is Ask Jev a chatbot like ChatGPT?
No. ChatGPT generates free text for you to read. Ask Jev returns a typed decision — YES, NO or IT DEPENDS — with a calibrated probability for each option, in a few hundred milliseconds. It decides; it doesn't converse.
Can Jev predict the future or give real advice?
No. Jev reads your question as written and returns calibrated odds about that wording — it's a decision model, not an oracle or an advisor. Treat the answer as a fast, structured opinion, not a fact or a recommendation to act on.
Why does it keep saying "it depends"?
Because Jev is calibrated to report honest uncertainty. Genuinely open questions get a spread-out distribution that lands on 'depends'; clear-cut questions snap to a confident YES or NO. The probabilities tell you how sure it is.
Is it really calling a live model?
Yes — every answer is a real, free Jev API call. The same endpoint that powers this toy is what developers use in production to route, classify, score and gate. Grab a key and see the TypeScript quickstart to build with it.