AI EngineerGuide

Convert OpenAI Decisions requests to System One (Jev, Clef)

The OpenAI request is preloaded below, converted to Jev. Paste your own to convert it, and read the warnings before you send it.

Runs in your browser. Nothing you paste is uploaded.

Converted request
Paste a request body, a structured-output schema, or a tool definition. Remove secrets first. Nothing leaves your browser.
curl
How the answers come back
QuestionOpenAIGeminiJev / Clef
Yes/no{type: "predicate", name, probability}{type: "noul", noul (0.01–0.99), confidence}{type: "noul", noul}
Pick one{type: "choice", name, choice, probabilities: [{value, probability}], confidence}{type: "choice", choice, probabilities: {option: p}, confidence}{type: "choice", choice, confidence, probabilities: {…}}
Ordered rubric{type: "score", name, score, probabilities: [{value, label, probability}], confidence}{type: "score", score, legend: {"0": …}, probabilities: {"0": p}, confidence}{type: "score", score, confidence, legend, probabilities}
Refusal{type: "refusal", name} (name is null for an unnamed question)not documentednot documented
Containerarray, in question ordermap keyed by question idmap keyed by question name (Clef wraps the response in result: unconfirmed)

17 warning types. Most link to the provider's docs; format facts come from the sources listed on each format pair page (last checked 2026-11-19).

What's different

OpenAI DecisionsTypeSafe Jev (System One)Cloudflare Clef (System One)
Endpointhttps://api.openai.com/v1/decisionshttps://api.typesafe.ai/v1/systemonehttps://api.cloudflare.com/client/v4/accounts/$CLOUDFLARE_ACCOUNT_ID/ai/run/@cf/cloudflare/clef
Modelgpt-6-lunajev-latestclef
QuestionsArray, in your order, 1–400Map keyed by question id, 1–? (max unconfirmed)Map keyed by question id, 1–64
Yes/nopredicatenoulnoul
Pick onechoicechoicechoice
Ordered rubricscorescorescore
Box (image)Not supportedboxNot supported
AuthAuthorization: Bearer $OPENAI_API_KEYAuthorization: Bearer $TYPESAFE_API_KEY (unconfirmed)Authorization: Bearer $CLOUDFLARE_API_TOKEN

Sources

Read 2026-11-19; the OpenAI pages re-read 2026-11-20.

  • OpenAI: “Ask 1–400 questions about the input.”
  • OpenAI: “Provide between 2 and 255 choices”
  • OpenAI: “Send requests to POST /v1/decisions using gpt-6-luna, the only currently supported model.”
  • OpenAI: “Use an input_json content part to pass structured evidence without first converting it to a string. Put the JSON object or array in data.”
  • OpenAI: Score levels: 2 to 10, from the API reference (read 2026-11-19); no verbatim line stored.
  • Jev: “A System One request has model, state and questions (a map); score criteria has maxItems: 10.”
  • Jev: Choice criteria: 1 to 255 options, from the System One request schema (read 2026-11-19). The maximum question count isn’t documented.
  • Clef: “Clef follows the System One API, so you can switch an existing Jev integration to Clef by changing the endpoint and model. Ask up to 64 questions per request.”
  • Clef: Model page for @cf/cloudflare/clef (and clef-flash). Option limits and the result response wrapper are unconfirmed here.

All formats · Classifier schema to decisions