Evaluate a state

Evaluate state with typed questions. Keys in answers match the question ids you sent. One request can mix all three types:

noul: yes or no. Returns the probability that the answer is yes, from 0 to 1.

choice: pick one option you defined, plus the full probability distribution. At most 255 options.

score: rate the state on an ordered rubric. The score is probability-weighted and can fall between levels. Use 2 to 10 levels.

instructions may be a string, object, or array. An object can hold the question in one field and data it refers to in others. Point at those fields with backticks, for example `potential_duplicate`.

Auth is a API token: Authorization: Bearer sk-.... Content-Type is application/json.

Provide your bearer token in the

Authorization

header when making requests to protected resources.

Example:

Authorization: Bearer ********************

Responses

🟢200Evaluation finished. `model` is the version that ran, which can differ from the requested alias. A request for `jev-latest` may return `jev-1.13.0`.

application/json

Evaluation finished. model is the version that ran, which can differ from the requested alias. A request for jev-latest may return jev-1.13.0.

Bodyapplication/json

🟠400The gateway rejected the body. Typical causes: missing `model`, `state`, or `questions`; `state` is not a string, object, or array; question `type` is not `noul`, `choice`, or `score`; `choice` criteria is not an object of 1 to 255 options; `score` criter

🟠401Missing or invalid API token. Check `Authorization: Bearer sk-...`

curl --location 'https://api.key4u.vn/v1/systemone' \ --header 'Authorization: Bearer <token>' \ --header 'Content-Type: application/json' \ --data '{ "model": "jev-latest", "questions": { "department": { "criteria": { "billing": "Payments, invoicing, refunds", "sales": "Pricing, upgrades, new accounts", "technical": "Bugs, outages, integrations" }, "instructions": "Which team should handle this?", "type": "choice" }, "frustration": { "criteria": [ "Calm", "Frustrated", "Very angry" ], "instructions": "How frustrated is the customer?", "type": "score" }, "is_urgent": { "criteria": { "false": "No urgency expressed", "true": "Explicitly time-sensitive" }, "instructions": "Does this convey urgency?", "type": "noul" } }, "state": "Help! My payouts have been failing for 3 days." }'

{ "answers": { "department": { "choice": "billing", "confidence": 0.81, "probabilities": { "billing": 0.88, "sales": 0, "technical": 0.12 }, "type": "choice" }, "frustration": { "confidence": 0.92, "legend": { "0": "Calm", "1": "Frustrated", "2": "Very angry" }, "probabilities": { "0": 0, "1": 0.95, "2": 0.05 }, "score": 1.05, "type": "score" }, "is_urgent": { "noul": 0.95, "type": "noul" } }, "model": "jev-1.13.0", "usage": { "input_tokens": 318, "output_tokens": 34 } }

Request

Responses