That is a LOT of code for a pretty basic feature.

> Never spend more than you budgeted

I assume that this is "$ spent on search + $ spent on LLM" < budget, but how do you handle the LLM spending more than you would expect on a request? Or is this handled by max_tokens and some form of pricing table? (and if so, how does caching play a role?)

https://github.com/tw93/Mole I see a certain conflict of interest.