Jev API: classification, routing and scoring

Put Jev's typed decisions inside your product. Ask up to 8 questions in one request and receive structured answers with probabilities.

One endpoint. Three question types.

Send text, JSON, or supported image inputs to /api/v1/systemone. Use Yes/No, Choice, and Score questions in the same request. Keep your API key on the server and validate the returned answer object before taking an action.

curl https://ai-jev.pro/api/v1/systemone \
  -H "Authorization: Bearer $JEV_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"state":"...", "questions":{"is_relevant":{"type":"noul","instructions":"Is this relevant?"}}}'

Response shape

Answers are keyed by the IDs in your request and include a typed value plus a provider-reported probability estimate. The values below illustrate the response shape only; they are not an evaluation result or accuracy benchmark.

{
  "answers": {
    "urgent": {
      "value": "Yes",
      "probability": 0.94
    },
    "topic": {
      "value": "Billing",
      "probability": 0.91
    }
  },
  "model": "jev-latest"
}

The Playground's initial output is also a static sample. Test representative, labeled examples and validate each category before using outputs to automate decisions. Probabilities are not independently calibrated; keep a human reviewer for consequential actions.

Request limits and errors

Models

Choose from the 19 listed models, including 18 enabled routes and Jev 1.14. Jev 1.13 remains available; model IDs and current input-token credit rates are in the catalog.