Docs
OfficialJev Practices & Patterns
State
How to shape what Jev evaluates: a string, a named object, or an array of messages. Every question in a request sees the same state and is evaluated independently.
- Format
- Docs
- Reading time
- 2 min
- Published by
- TypeSafe AI
- Last verified
- Sep 2026
Works forJev for Scoring
Used inJev for Software & DevTools
Highlights
Quickstart
json
{
"ticket": {
"subject": "Duplicate charge",
"messages": [
{"from": "customer", "text": "I was charged twice for order A-104. Please refund the duplicate."},
{"from": "support", "text": "We are checking the charges."}
]
}
}Watch out
Jev accepts text only, and packing in large amounts of unrelated detail causes context rot, so trim state to what the question needs.
Questions about this entry
Short answers drawn from the linked project or page.
The page recommends objects for most requests so every part of the state has a descriptive name, and suggests a plain string for a single message or passage. An array models a message sequence with roles.
Yes. One request pairs one state with one or more questions and can mix Choice, Score, and Noul freely; every question sees the same state and is evaluated independently. Trim unrelated detail to avoid context rot.
Jev accepts text only: a string, a JSON object, or an array of text values. Images, audio, and video are not supported yet, so any multimedia has to be transcribed or described to text before it goes in the state field.
Jev's primary training language is English. Other languages, including CJK scripts, are accepted but currently run at lower accuracy, so calibrate or route non-English states through review.
State is passed in the state field of a request, alongside the questions you want answered. In Python you hand the string, dictionary, or list straight to client.system_one(state=...).
Put related information in one state, because that state is the material the model judges. An object can hold a conversation, an order, and a refund policy at once, and every question sees all of it.
The state carries the content and supporting facts; the questions define the judgments to make about it. Keep a refund request and policy in the state, then ask whether a refund was requested and whether policy supports it.
More Jev resources like this
A maintained list of jev-1.13's known failure modes, literal reading, unreliable counting, dates as text, indirection, context rot, and contradictory criteria, each with a guardrail.
Jev Practices & PatternsDocs#official#docs#evaluation
Official
The experimental image-input reference: PNG or JPEG files in a top-level assets map, referenced by string anywhere in state, instructions, or criteria, behind the TypeSafe-Danger-Zone-Images header.
Jev Practices & PatternsDocs#official#docs#vision
Official
The box question type: one bounding box per question, returned as four edges that are fractions of the image, plus confidence, with no criteria field and no not-found value.
Jev Practices & PatternsDocs#official#docs#vision
Official
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
Rate content against 2 to 10 ordered levels. Returns a probability-weighted score that can land between levels, per-level probabilities, and confidence, with the arithmetic left to code.
Jev Practices & PatternsDocs#official#primitives#score
Official
Instructions, option descriptions, score levels, and noul criteria all accept JSON, so you can label question parts, pass schemas and taxonomies, and keep candidate paths alive.
Jev Practices & PatternsDocs#official#primitives#architecture
Official
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.