Decisions

title: Decisions description: Classify, route and score content with decision models through POST /decisions on api14. Typed questions in, calibrated probabilities out. keywords: decisions API, /decisions, decision model, classification API, AI scoring API, yes/no probability, routing model, rubric scoring

Decision models read a piece of content (the state) and answer typed questions about it with probabilities instead of free text. Use them for routing, moderation, triage and rubric scoring, where you need a predictable, machine-readable answer.

POST https://api14.com/api/v1/decisions

Decision models do not accept /chat/completions. The official OpenAI SDKs have no decisions method, so call it with plain HTTP or the SDK's generic post (every example below does).

Request

Question types

Every question has type, instructions and criteria.

Example

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api14.com/api/v1",
  apiKey: process.env.API14_API_KEY,
});

const result = await client.post("/decisions", { body: {
  "model": "liquid/d1",
  "state": "Customer: I was charged twice for my order and want my money back.",
  "questions": {
    "team": {
      "type": "choice",
      "instructions": "Which team should handle this message?",
      "criteria": {
        "billing": "Payments, refunds, invoices",
        "support": "Product questions and bugs",
        "sales": "Pricing and new purchases"
      }
    },
    "urgent": {
      "type": "noul",
      "instructions": "Does this need a reply today?",
      "criteria": {
        "true": "Money, outage or legal risk",
        "false": "Can wait"
      }
    }
  }
} });
console.log(result.answers);

curl https://api14.com/api/v1/decisions \
  -H "Authorization: Bearer sk-api14-..." \
  -H "Content-Type: application/json" \
  -d '{
    "model": "liquid/d1",
    "state": "Customer: I was charged twice for my order and want my money back.",
    "questions": {
      "team": {
        "type": "choice",
        "instructions": "Which team should handle this message?",
        "criteria": { "billing": "Payments, refunds, invoices", "support": "Product questions and bugs", "sales": "Pricing and new purchases" }
      },
      "urgent": {
        "type": "noul",
        "instructions": "Does this need a reply today?",
        "criteria": { "true": "Money, outage or legal risk", "false": "Can wait" }
      }
    }
  }'

Response

{
  "id": "gen-dec-1791900000-a1",
  "model": "liquid/d1",
  "answers": {
    "team": { "type": "choice", "choice": "billing", "confidence": 0.93, "probabilities": { "billing": 0.93, "support": 0.05, "sales": 0.02 } },
    "urgent": { "type": "noul", "noul": 0.81 }
  },
  "usage": { "input_tokens": 142, "output_tokens": 0, "cost": 0.0000071 }
}

Log the model string with each answer: an alias such as ~typesafe/jev-latest can point to a newer version later.

Pricing

Decision models are billed per input token at the model's listed price. Output tokens are free. The cost is in usage.cost and debited from your USD balance.

Listing decision models

curl "https://api14.com/api/v1/models?category=decisions"

Errors

See Errors.