pipecat-ai/voice-ui-kit ? reverse-engineered prompt

Build me a React voice AI UI kit for Pipecat that makes it easy to put together a working voice app fast.

I want reusable components, hooks, and a few ready to use template screens so I can quickly spin up things like a chat style voice console, a simple control bar, audio controls, a visualizer, error states, loading states, and a connect and disconnect flow. It should feel polished out of the box, work well on desktop and mobile, and be easy to customize with theme colors and CSS variables. Please include a full screen template app, a debug console for testing and benchmarking, and a few example apps that show how to build a custom UI on top of the kit.

If you need to check current Pipecat or transport docs while wiring it up, go ahead and look them up online.