Hey!
llama-cpp-python doesn’t come with a built-in web UI by default — so there’s no traditional “UI file” in the repo. But if you're using something like llama-cpp-server or any gradio-based wrappers (like KoboldCPP / Oobabooga), the UI layer is added by that wrapper.

If you’re building your own, the button logic would go in the frontend JS/HTML if it's a browser UI, or in the Python web layer if you're using FastAPI or Gradio.

What are you using to serve the model? Happy to help you locate the right spot.
(I’ve built some logic engines on top of this stack too.)

GitHub patrol never stops 👣