Jev, a decision layer between your agent and its actions
Jev is a hosted decision model from TypeSafe AI. It does not write text: it returns typed answers with probabilities, such as one choice from a list, a score or a yes-or-no probability. We use it to filter, route and guard what an agent is about to do before it does it.
Information page. Buying opens once the price of this add-on is set.
What a decision model is
A language model writes text; a decision model answers a narrow question with a typed result your code can use directly. Jev offers three kinds of answer.
- Choice
- Picks one option from a defined set, for example which team should handle a request.
- Score
- Places something on a defined scale, for example how urgent a message is.
- Yes or no
- Returns the probability that a condition holds, for example whether a message asks for a refund.
What we would use it for
- Hold an outgoing email for approval when it looks like a commitment or a complaint.
- Route an incoming request to the right agent or person.
- Block a deletion or a payment request unless a person confirms it.
How access works
You use your own TypeSafe AI account and key; we connect it to your agent and write the questions with you. Jev currently accepts text only.
Common questions
Is Jev another chat model?
No. It returns typed decisions and probabilities, not generated text.
Who pays for Jev?
You do, through your own TypeSafe AI account. Our add-on covers the integration work.
Are you affiliated with TypeSafe AI?
No. OpenClawPro is an independent service.