This is what I want!

The most time consuming part of coding with an LLM is collecting the right context to put in the prompt. But too much context can be just as bad. I will look into replit. 🙏