[DJC_Analytics]

This is really interesting. Have you thought of making something similar for non-coders? I imagine that it shouldn't be too different as the same problem occurs in all sufficiently long chats with an LLM.

But, if this acts as a KB as I seem to understand it does from a cursory look, this could be a much bigger tool for a much larger audience.

Thoughts?

[cagangedik]

Thanks! The problem isn’t AI chat, it’s tracking what the AI actually does to files.

Since we rely on MCP, starting with environments like GitHub and IDEs makes this tractable.

If we can own more of that surface (i.e. build our own tool), expanding beyond devs is definitely on the roadmap.