duixcom/Duix-Avatar ? reverse-engineered prompt

Build me a Windows desktop app for creating an AI avatar that can clone a person’s face and voice, then generate talking videos completely offline. I want a simple, beginner friendly interface where someone can add their own model, upload a photo or sample video, record or paste in text, and then render a video with natural lip sync and speech.

It should work locally on Windows, with Docker handling the backend services, and the app should guide the user through setup if anything is missing. Please make it feel polished and easy to use, with clear status messages, progress while models are loading or videos are generating, and basic support for managing multiple models. If you need to look up current docs online while wiring things up, go ahead.