I don't really see how you're comparing Pydantic AI here to Typescript. I'm assuming you meant simply Pydantic.

Just comparing an agent framework written in python (with focus on being "typesafe") to one (any) written in typescript