Full guide: Choice, Score and Noul: the three primitives
In practice a Noul is a named question in the request with its type set to noul. The docs’ example sends a support message as the state and asks {"type":"noul","instructions":"Is the customer asking for a human agent?"}. The answer comes back as {"type":"noul","noul":0.99}: a 99% probability that the answer is yes. There is no special Noul data type in the response. The value is a plain number, and your code compares it against a threshold like any other float.
Reading the number is direct. A value near 1 is a strong yes and a value near 0 is a strong no. Near 0.5 the model gives both answers similar probability.
Gating a Noul means picking thresholds on the probability itself, since there is no confidence field. A screening check might act above 0.9, ignore below 0.2, and send the band in between to review. That is a different setup from Choice and Score, where the answer and the confidence in it are separate numbers you can combine. Extraction models such as Clairic take typed fields instead of questions; see extraction model.
Where the word “Noul” comes from is not documented. TypeSafe’s primitive page defines the behaviour and says nothing about the origin of the term, so any explanation you find elsewhere is a guess. It has nothing to do with the English dictionary word “noul”, a variant of the obsolete word “noll”, meaning the head.
Noul suits binary screens that run on every request, such as whether a message contains a prompt injection attempt or whether a retrieved passage supports the claim it is cited for. Choice, Score and Noul puts all three question types side by side with full request and response examples, and covers when a Noul beats a two-option Choice.