Decisions

Mercury Decide evaluates named questions against a shared state and returns typed answers with probabilities. Use it for routing, classification, and scoring with POST /v1/decisions and the model mercury-decide.

Send a decision request

Set INCEPTION_API_KEY to your API key. Each request includes a model, a state, and a non-empty map of questions. The names you choose for questions are preserved in the response.

curl https://api.inceptionlabs.ai/v1/decisions \
  -H "Authorization: Bearer $INCEPTION_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "mercury-decide",
    "state": {"message": "Please refund the duplicate charge on my account."},
    "questions": {
      "billing": {
        "type": "noul",
        "instructions": "Is this request about billing?"
      },
      "team": {
        "type": "choice",
        "criteria": {
          "billing": "Payments, charges, and refunds",
          "support": "Technical support and other requests"
        }
      },
      "urgency": {
        "type": "score",
        "instructions": "How urgent is this request?",
        "criteria": ["Can wait", "Needs attention soon", "Needs immediate attention"]
      }
    }
  }'

import json
import os
import urllib.request

payload = {
    "model": "mercury-decide",
    "state": {"message": "Please refund the duplicate charge on my account."},
    "questions": {
        "billing": {"type": "noul", "instructions": "Is this request about billing?"},
        "team": {
            "type": "choice",
            "criteria": {
                "billing": "Payments, charges, and refunds",
                "support": "Technical support and other requests",
            },
        },
        "urgency": {
            "type": "score",
            "instructions": "How urgent is this request?",
            "criteria": ["Can wait", "Needs attention soon", "Needs immediate attention"],
        },
    },
}
request = urllib.request.Request(
    "https://api.inceptionlabs.ai/v1/decisions",
    data=json.dumps(payload).encode(),
    headers={
        "Authorization": f"Bearer {os.environ['INCEPTION_API_KEY']}",
        "Content-Type": "application/json",
    },
    method="POST",
)
with urllib.request.urlopen(request, timeout=60) as response:
    decision = json.load(response)
print(decision["answers"])

const response = await fetch('https://api.inceptionlabs.ai/v1/decisions', {
  method: 'POST',
  headers: {
    Authorization: `Bearer ${process.env.INCEPTION_API_KEY}`,
    'Content-Type': 'application/json',
  },
  body: JSON.stringify({
    model: 'mercury-decide',
    state: { message: 'Please refund the duplicate charge on my account.' },
    questions: {
      billing: { type: 'noul', instructions: 'Is this request about billing?' },
      team: {
        type: 'choice',
        criteria: {
          billing: 'Payments, charges, and refunds',
          support: 'Technical support and other requests',
        },
      },
      urgency: {
        type: 'score',
        instructions: 'How urgent is this request?',
        criteria: ['Can wait', 'Needs attention soon', 'Needs immediate attention'],
      },
    },
  }),
});
if (!response.ok) {
  throw new Error(`Decision request failed (${response.status}): ${await response.text()}`);
}
const decision = await response.json();
console.log(decision.answers);

Question types

Type	Criteria	Answer
`noul`	Optional `true` and `false` descriptions	`noul`: the probability of a positive answer, from 0 to 1
`choice`	A map of 1–255 option names to descriptions; descriptions may be `null`	`choice`: the highest-probability option, plus `probabilities` for every option and `confidence`
`score`	An ordered array of 1–10 level descriptions, starting at level 0	`score`: the probability-weighted average of the level indices, plus `legend`, `probabilities`, and `confidence`

The state, instructions, and criterion descriptions can be strings, JSON objects, or arrays. Instructions may be omitted or null. Questions are evaluated independently against the shared state. A score can be fractional; it is not restricted to a single level.

Read the response

The response contains the model name, an answers map keyed by your question names, and usage. For example, a request with a single noul question named billing could return:

{
  "model": "mercury-decide",
  "answers": {
    "billing": {"type": "noul", "noul": 0.97}
  },
  "usage": {
    "input_tokens": 48,
    "output_tokens": 1
  }
}

The values above are illustrative. For choice, probabilities uses your option names. For score, it uses string indices such as "0" and "1", and legend maps those indices back to your criteria. Both include a confidence value between 0 and 1.

Pricing and limits

Mercury Decide costs $0.04 per million input tokens. Only input tokens are billed; output tokens are reported in usage.output_tokens but are not charged. See Models & Pricing.

Send requests to /v1/decisions. Mercury Decide does not support /v1/chat/completions, streaming, or tool calling.
Use Error Codes for authentication, rate-limit, and service errors.

Getting Started

API reference

Capabilities

Guides

Resources

Support

Send a decision request

Question types

Read the response

Pricing and limits

​Send a decision request

​Question types

​Read the response

​Pricing and limits

Send a decision request

Question types

Read the response

Pricing and limits