Typed questions, calibrated probabilities, one call.

Ask typed questions — which option, yes or no, how much — and get a calibrated probability for every answer, in one request. Text, images and video. Free to use.

Access

The API answers only under the base URL issued to you — path included. There is no public sign-up, and every other path returns a plain 404.

Your first call

curl -s "$BASE_URL/v1/decisions" \
  -H "content-type: application/json" -d '{
  "model": "vrd/decision-1",
  "state": "The checkout page returns 500 errors after the latest deploy.",
  "questions": {
    "route": {
      "type": "choice",
      "instructions": "Which team should handle this ticket?",
      "criteria": {
        "billing": "Payments, invoices and refunds",
        "engineering": "Outages, errors and bugs",
        "sales": "Plans, pricing and contracts"
      }
    },
    "urgent": { "type": "noul", "instructions": "Does this need attention today?" }
  }
}'

Models

Model	Profile
`vrd/decision-1`	flagship

Pick the smallest model that meets your accuracy bar; larger models are slower under load.

No per-key limits

There are no quotas on keys and nothing to configure. Send requests without a key; the budget is 6.9M input tokens per minute in total — effectively no ceiling for a normal client — and the answer to exceeding it is a 429 with retry-after. Details.