Hey, I tried it out - it's pretty cool. I haven't sprung for GPT4, so I tried really hard to get the tokens under 4k, but never made it.

Despite that, the output of this was spot on. If this could utilize the api, it would be even better. I could see this tool eventually being able to keep track of what code is in what file, spitting that out to GPT4, and sort of being a liaison between the codebase and GPT.