Jev & TypeSafe
What Jev is, how System One fits your stack, and where this directory ends and TypeSafe docs begin.
What is Jev?
Jev is TypeSafe's flagship System One model. You send structured state and typed questions; the API returns discrete answers with per-option probabilities and confidence scores. Your application code applies thresholds and executes side effects.
Where to start
Read the introduction at typesafe.ai, follow the quick start on docs.typesafe.ai, and poke the browser playground. For production, install @typesafe-ai/sdk or typesafe-sdk on Python and call POST /v1/systemone.
Relationship to this directory
jev.aitools.fyi is an independent curated directory by aitools.fyi. Listings link to third-party projects; only official TypeSafe properties are authored by TypeSafe AI.
Choice, Score, and Noul
Choice picks one label from a fixed set with explicit probabilities. Score returns a numeric rating with uncertainty you can map to bands. Noul answers structured true or false style questions with confidence. Compose these primitives instead of prompting for free text and regex-parsing the reply.
Request shape in practice
Production clients send JSON state plus an array of typed questions in one POST to /v1/systemone. The response carries aligned answers with per-question confidence. Your service layer applies thresholds, logs decisions, and triggers side effects. There is no hidden chain-of-thought stream to consume.
When Jev is the wrong tool
Skip Jev when you need long-form copy, open-ended brainstorming, or narrative reasoning. Reach for an LLM or a template engine there. Many teams pair both: Jev for gates and routing, generative models for user-visible language.
Know someone still labeling with chat?
Send them this guide. Low spam, high signal.
Drop it in Slack. Tag TypeSafe if it saves someone a sprint.
More context on Jev Directory and the llms.txt map for AI crawlers.