Docs
OfficialJev Guides & Articles
TypeSafe Documentation
The canonical introduction: Jev is TypeSafe's flagship model and the first System One model. Send state and typed questions, get structured answers your code can use directly.
- Format
- Docs
- Reading time
- 2 min
- Published by
- TypeSafe AI
- Last verified
- Sep 2026
Works forJev for Scoring
Used inJev for Software & DevTools
Highlights
- Jev is TypeSafe's flagship model and the first System One model, returning typed values and probability distributions instead of generated text.
- Choice and Score answers carry a 0-1 confidence; Noul returns only the probability of yes, with no separate confidence field.
- All three question types mix in one call, and each question is evaluated in parallel and in isolation against the same state.
- Adding questions barely changes response time and does not create context rot because each question is evaluated independently.
- The docs recommend decomposing broad judgments like 'rate this startup pitch' into atomic questions and weighting answers in code.
Watch out
Judgments that need extended reasoning should be decomposed into separate questions whose answers are combined with logic in your code.
Questions about this entry
Short answers drawn from the linked project or page.
Choice and Score. Both return a 0-1 confidence alongside probabilities, while Noul returns only the probability of yes with no separate confidence field, because the probability is the answer.
No. Each question is evaluated in parallel and in isolation against the same state, so adding questions barely changes response time. Broad judgments are decomposed into atomic questions whose answers are weighted in code.
A model built to make fast, structured decisions that software can use directly, returning typed values and probability distributions instead of text for a human to read.
Choice returns choice, probabilities and confidence; Score returns score, probabilities and confidence; Noul returns a 0-1 probability of yes.
Decompose it. Ask each factor as a separate atomic question and combine the results with your own logic or weights in code.
Yes. All three question types can be mixed in one API call, and every question is evaluated in parallel and in isolation against the same state.
No text generation and no parsing. An LLM produces prose you must coerce into structured decisions and parse back, while Jev returns typed values and probability distributions your code can branch on directly.
More Jev resources like this
The raw HTTP contract behind every SDK: POST a state plus typed noul, choice, and score questions to /v1/systemone, and get one answer per question, with error codes and retry guidance.
Jev Guides & ArticlesDocs#official#docs#api
Official
The idea behind Jev: models that return typed decisions and probabilities instead of generated text, named after Kahneman's fast-thinking System 1, with confidence to route around uncertainty.
Jev Guides & ArticlesDocs#official#docs#architecture
Official
Guidance on how Jev fits coding agents: it is not a drop-in agent model, so the page routes readers to routing, rubric scoring, and truth checks instead.
Jev Guides & ArticlesDocs#official#docs#coding-agent
Official
Everything you need to make your first request: install an SDK, get an API key, and send state plus typed questions.
Jev Guides & ArticlesDocs#docs#official#quickstart
Official
The Jev Day release post: Jev 1.14.0 under jev-preview plus three danger-zone features — composable images on both models, the box question type on 1.14 only, and opt-in 262k context.
Jev Guides & ArticlesArticle#official#release#vision
Official
OpenAI's API documentation page for Decisions, a hosted endpoint that returns bounded answers rather than generated text; the captured page is its header, pointing to the docs index at llms.txt.
Jev Guides & ArticlesDocs#community#openai#docs
Community
From the Jev guides
Original write-ups on what this entry is for.
From the Jev AI model community
Posts from builders shipping with Jev right now.