afngh/dynamo ? reverse-engineered prompt

Build me a simple web app where I can chat with the model in real time.

I want a clean React frontend with a chat screen, a message composer, and a settings panel for things like temperature and max tokens. When I send a prompt, the reply should stream in token by token instead of waiting for the full answer. Make it feel smooth and modern, with clear loading states, error handling, and an easy way to start a new chat.

Hook it up to the Django backend that serves the streaming response, and make sure the frontend can read the live text as it comes in. If there are any setup details or API quirks, check the current docs online if you need to. Keep the UI polished, responsive, and ready to run locally with a simple setup.