Call decision models - Domino

Decision models such as TypeSafe’s Jev don’t generate text. You send a piece of state and a set of typed questions, and the model returns a typed answer to each, with calibrated probabilities. Use them for routing, classification, gating, and guardrail checks where you need a fast, cheap, machine-readable answer.

Question types

Type	Returns
`noul`	The probability that a statement is true.
`choice`	One of up to 255 options.
`score`	A position on an ordered scale of 2 to 10 levels.

Add a provider

Provider type	Use it for	Base URL
TypeSafe (Jev)	TypeSafe’s own API	Leave blank
System One compatible	Any other host of the same protocol, including OpenRouter, Vercel AI Gateway, Codiv, self-hosted OpenJev, and LiteLLM	The host’s base URL, without `/v1`

Model names differ by host, for example jev-latest on TypeSafe and typesafe/jev-1.13 on OpenRouter. The built-in price book covers the Jev models.

Call a decision alias

Send POST /v1/systemone with a decision alias as model. The TypeSafe software development kit (SDK) works if you set its baseURL to the gateway root. In the Playground, picking a decision alias opens a decision editor where you build questions, see answers as probability distributions, and copy Python, TypeScript, or cURL snippets.

What applies to decision aliases

Grants, budgets, bring your own token (BYOT) grants, tag headers, input guardrails on state, caching, and retries all apply. Streaming, fallback chains, and output guardrails don’t. A decision alias can’t serve chat, responses, embeddings, or Claude Code, and returns a 400 if you try.

Control access, spend, and usage: grant access to decision aliases and set budgets.
Configure guardrails and fallback routing: screen the state you send.

Last modified on October 15, 2026

⌘I