xinnan-tech/xiaozhi-esp32-server ? reverse-engineered prompt

I want to build a self-hosted server for a DIY smart AI assistant called "xiaozhi" that runs on an ESP32 device. The main goal is to create a backend that can process voice commands from the hardware, connect to an AI model for intelligent conversations, and then generate spoken responses.

The assistant should be able to do practical things like control home appliances, play music, or give me a weather update. I've also seen it can have more advanced features like recognizing different users by their voice and even identifying objects if a camera is connected, so please include those capabilities.

I'll also need a web-based control panel to manage the system, configure API keys, and see what's happening. Please set up the entire project to be easily deployed using Docker, as that seems like the simplest way to run it. For the AI services like speech recognition and synthesis, please configure them to use free-tier or open-source options by default.