Jev AI

Jev is a decision model from TypeSafe. You give it some text and a typed question; it gives you back an answer and the probability behind it. It does not write prose, and it cannot invent an option you did not offer.

Try it in your browser What it costs

Jev AI is an independent site. It is not TypeSafe, and TypeSafe has not reviewed or endorsed anything on this page.

How it differs from an LLM

A language model generates one token at a time and you parse the result. Jev does not generate. It evaluates your input once and returns a typed value: one of your options, a number on your scale, or a probability between 0 and 1. There is no JSON to parse and no schema to validate against.

The practical consequence is latency and price. TypeSafe’s own material quotes roughly 70 to 500 milliseconds — that number is server-side processing time and does not include network transit. We measured the end-to-end round trip ourselves on 2026-09-20: 20 calls, made from mainland China, routed through OpenRouter (not a direct connection to TypeSafe), using a customer-support-triage payload (one choice question, one score question, one noul question). Fastest call: 309 ms. Median: 423 ms. 90th percentile: 1041 ms. Slowest: 1077 ms. Twenty calls from one location is a small sample, and a connection closer to TypeSafe’s infrastructure would likely be faster — but end-to-end, from a real browser, is the number that actually matters to an integrator, and it is well above the server-side figure. Input costs $0.042 per million tokens with output billed at zero.

The three question types

choice — pick one of the options you defined. Returns the pick, a probability for every option, and a confidence value.
score — place the input on an ordered scale you describe in words. Returns a fractional score, so 1.81 on a three-point scale means it sits most of the way toward the top.
noul — a single yes/no judgement, returned as a probability from 0 to 1.

You can ask several questions in one request, and they are evaluated together.

How accurate is it, really

On TypeSafe’s own four-workflow benchmark, Jev averaged 67.8% agreement with the reference answers. GPT-5.6 Terra scored 67.9%, Opus 5 scored 73.1%, and GPT-5.6 Sol scored 74.1%. On invoice processing specifically Jev managed 61.8% against Sol’s 79.1%. Its best workflow was customer service, at 76.0%.

Two things follow from that. First, Jev is not more accurate than a language model on these tasks — it is faster and cheaper, and on some tasks meaningfully less accurate. Second, the benchmark’s reference answers were produced by averaging the responses of GPT-6 Astra and Claude Fable 5.1. No human labelled anything. So 67.8% measures agreement with two frontier models, not correctness.

Where to go next

Playground — run it on your own text with ready-made templates, then export the request as code.
Pricing and access — what it costs against the small models you would otherwise use, and which route to call it through.