Jev developer tools

Build a text classifier with a bounded answer set.

Test a custom category set, inspect the result, and generate starter TypeScript or Python code from the same configuration.

JevClassify is an independent tool and is not affiliated with, sponsored by, or endorsed by TypeSafe AI.

Quick answer

What is Jev?

Jev is used for constrained software decisions: a shared input and a typed question go in, then a bounded choice, score, or boolean result comes out. JevClassify uses its Choice flow for one-label text classification.

Input

Text, a classification instruction, and two to ten custom categories.

Output

One selected category and probabilities only when the provider returns them.

Boundary

Results can be wrong and are not a safe substitute for review in consequential workflows.

Implementation source: Jev is called through Vercel AI Gateway using the verified evaluation-model flow. Generated code never includes the submitted text or a real API key.

State + typed question

Make the answer space explicit.

State“I was charged twice.”
QuestionWhich support category fits best?
ChoicesBilling · Technical · Sales · Other
ResultRender only the fields returned by the actual provider.
AI TEXT CLASSIFIER2–10 labels
Categories

AI processing: Text submitted for classification may be sent to third-party AI services. Do not submit passwords, API keys, payment card data, health information, or other sensitive information. Privacy

Developer paths

Take the configuration into your app.

Jev FAQ

Questions before you integrate.

What is Jev?

Jev is a model for bounded, typed decisions in software. It evaluates text or other state against a finite set of choices and returns the selected answer with provider-supported probability fields.

Does JevClassify generate free-form text?

No. JevClassify uses Jev for classification and generates deterministic integration templates from the selected classifier configuration.

Is JevClassify an official TypeSafe AI product?

No. JevClassify is an independent tool and is not affiliated with, sponsored by, or endorsed by TypeSafe AI.