TypeSafe AI

San Francisco lab behind Jev, a model that returns typed decisions instead of text.

TypeSafe AI is a San Francisco AI lab building what it calls System One models, made to be called by software rather than chatted with. Jev, its first, reads a state given as text and answers yes/no, multiple-choice and scale questions about it with typed values and calibrated probabilities. It generates no text, code or explanations.