Re-creating the Demo

[buddha314]

Hello! I'm using the prompts provided to accompany the demo video. I've picked Claude 3.7 as my LLM. When I run the prompt, I get pretty uneven behavior from the agents. I've tried re-doing it several times to get a feel of what's going on.

• The Coder often fails to write the files at all
• The Agent isn't installed in the front end
• The AI Chat isn't enabled in the new IDE

I expect some variance with LLMs, of course, but I'm wondering if there is a way to fine tune the prompts to get more consistent results.