FAQ
Jev fit, verdicts, cost, sharing.
What is Jev
Jev is TypeSafe's calibrated decision engine for semantic judgment: Choice picks a category from a fixed set, Score rates along ordered levels, and Noul answers yes/no with a confidence your code acts on. Latency is about 100ms and cost is $0.042 per million input tokens with free output.
Choice vs Score vs Noul
Use Choice when the answer is one of a fixed set of categories, like intent, component, or tool. Use Score when the answer is an ordered level, like urgency P0–P3 or fit 1–5. Use Noul when the answer gates action, like needs-human, has-PII, or injection-detected. Combine them per pattern.
When not to use Jev
Do not use Jev for prose generation, multi-step reasoning, or exact precision such as math, dates, or regex. Send those to an LLM or keep them in code. You send judgment to Jev, compose the verdict in code, and call the LLM only where prose is the product.
Questions
- What is FitForJev?
- You paste a fuzzy AI idea. You get the typed questions TypeSafe Jev answers, plus a fit verdict computed in code. Paste a problem and you get a verdict (native, hybrid, weak, not_jev, use_code, or need_info), the Jev questions inside it, SDK code, and a TypeSafe Playground link.
- How is the verdict computed?
- TypeScript computes it. You answer ten literal rubric questions in parallel. Five signals add to the score, three penalties subtract, and computeFit composes your answers into a verdict. Jev never judges its own fit; you own the verdict logic in code.
- What do the verdict classes mean?
- Native is a pure Jev decision layer. Hybrid means Jev decides and an LLM writes. Weak fit has a small salvageable Jev core. Not Jev needs a different tool. Use Code means regex or math suffices. Need Info means the problem is too vague to assess.
- When should I not use Jev?
- Do not use Jev for prose generation, multi-step reasoning, or exact precision such as math, dates, or regex. Send judgment to Jev at 100ms, keep deterministic rules in code, and call the frontier model only where prose is the product.
- What does Jev cost compared to an LLM?
- Per 100k calls, Jev costs about $0.004 at 100ms latency versus about $0.50 at 2,500ms for a frontier LLM. See the economics table on the homepage and the TypeSafe docs for current pricing.
- How do I share my verdict?
- Each evaluation gives you a share link under /s/. The verdict renders server-side, so X, LinkedIn, and Slack unfurl it with title and image. Share pages carry noindex so they stay out of the search index.
Evaluate your own problem or browse the pattern catalog.