This is a great use case for AI — turning vague intent into structured instructions. The cooking domain is nice because the feedback loop is fast (you know within an hour if it worked). What model are you using under the hood?

[dead]