Interpret supplied input, then execute a model

Supply the complete aityx model and evidence. Valid explicit facts are preserved; OpenAI establishes unresolved fields and their confidence from the supplied data. Extracted values are checked against the schema before local evaluation. Missing required evidence fails instead of being guessed. No model is created and Jev is not called. Interpretation and execution settle one System Two charge using API input/output tokens. Provider usage is reported separately, not charged again.

POST/v1/systemtwo/decisions

Authorization

AuthorizationBearer token · headerrequired

Account API key

Request body

requiredapplication/json

decisionNativeModelrequired

Complete native decision definition. Answer names and types are inferred locally from its decisions; no questions or output metadata is required.

Show properties

namestringrequired

inputsobjectrequired

Inputs declare native types. In combined execution, from="jev:<JSON Pointer>" binds a field or complete answer from the original Jev result. Objects and lists support typed records or Any.

decisionsobjectrequired

functionsobject

typesobject

stateStaterequired

Show properties

One of:

object

object

any[]

Array of any

any

string

string

Responses

200

Interpreted facts, answers, inline receipt and one settled System Two charge

answersobjectrequired

Named Jev and native outputs. Native decisions own overlapping names; original Jev facts remain in the input trace.

receiptReceiptrequired

Computed values and input provenance without repeating supplied state. Not persisted.

Show properties

readobject

Jev scalar facts and bound response data (source=jev with optional from pointer) or OpenAI observations (source=openai).

provenanceobject

decisionsobject

functionsobject

usageUsagerequired

Show properties

input_tokensintegerrequired

min 0

output_tokensintegerrequired

min 0

decision_errorDecisionError

Show properties

codestringrequired

detailstringrequired

errorsobject[]

400

Request failed; no completed operation charge is settled. Completed Jev with model failure uses the HTTP 200 partial response instead.

statusinteger

codestring

detailstring

errorsobject[]

401

Request failed; no completed operation charge is settled. Completed Jev with model failure uses the HTTP 200 partial response instead.

statusinteger

codestring

detailstring

errorsobject[]

402

insufficient_balance; available funds cannot cover the request reservation

statusinteger

codestring

detailstring

errorsobject[]

403

Request failed; no completed operation charge is settled. Completed Jev with model failure uses the HTTP 200 partial response instead.

statusinteger

codestring

detailstring

errorsobject[]

422

Request failed; no completed operation charge is settled. Completed Jev with model failure uses the HTTP 200 partial response instead.

statusinteger

codestring

detailstring

errorsobject[]

502

Request failed; no completed operation charge is settled. Completed Jev with model failure uses the HTTP 200 partial response instead.

statusinteger

codestring

detailstring

errorsobject[]

503

Request failed; no completed operation charge is settled. Completed Jev with model failure uses the HTTP 200 partial response instead.

statusinteger

codestring

detailstring

errorsobject[]

504

Request failed; no completed operation charge is settled. Completed Jev with model failure uses the HTTP 200 partial response instead.

statusinteger

codestring

detailstring

errorsobject[]

Request

curl -X POST 'https://api.aityx.ai/v1/systemtwo/decisions' \
  -H 'Authorization: Bearer YOUR_TOKEN' \
  -H 'Content-Type: application/json' \
  -d '{
  "decision": {
    "name": "string",
    "inputs": {},
    "decisions": {},
    "functions": {},
    "types": {}
  },
  "state": {}
}'

const response = await fetch("https://api.aityx.ai/v1/systemtwo/decisions", {
  method: "POST",
  headers: {
    "Authorization": "Bearer YOUR_TOKEN",
    "Content-Type": "application/json"
  },
  body: "{\n  \"decision\": {\n    \"name\": \"string\",\n    \"inputs\": {},\n    \"decisions\": {},\n    \"functions\": {},\n    \"types\": {}\n  },\n  \"state\": {}\n}"
});

import requests

response = requests.post(
    "https://api.aityx.ai/v1/systemtwo/decisions",
    headers={
        "Authorization": "Bearer YOUR_TOKEN",
        "Content-Type": "application/json"
    },
    data="{\n  \"decision\": {\n    \"name\": \"string\",\n    \"inputs\": {},\n    \"decisions\": {},\n    \"functions\": {},\n    \"types\": {}\n  },\n  \"state\": {}\n}",
)

Response

{
  "answers": {},
  "receipt": {
    "read": {},
    "provenance": {
      "property1": "state",
      "property2": "state"
    },
    "decisions": {},
    "functions": {}
  },
  "usage": {
    "input_tokens": 0,
    "output_tokens": 0
  },
  "decision_error": {
    "code": "string",
    "detail": "string",
    "errors": [
      {}
    ]
  }
}

{
  "status": 0,
  "code": "insufficient_balance",
  "detail": "string",
  "errors": [
    {}
  ]
}

{
  "status": 0,
  "code": "insufficient_balance",
  "detail": "string",
  "errors": [
    {}
  ]
}

{
  "status": 0,
  "code": "insufficient_balance",
  "detail": "string",
  "errors": [
    {}
  ]
}

{
  "status": 0,
  "code": "insufficient_balance",
  "detail": "string",
  "errors": [
    {}
  ]
}

{
  "status": 0,
  "code": "insufficient_balance",
  "detail": "string",
  "errors": [
    {}
  ]
}

{
  "status": 0,
  "code": "insufficient_balance",
  "detail": "string",
  "errors": [
    {}
  ]
}

{
  "status": 0,
  "code": "insufficient_balance",
  "detail": "string",
  "errors": [
    {}
  ]
}

{
  "status": 0,
  "code": "insufficient_balance",
  "detail": "string",
  "errors": [
    {}
  ]
}