iperov/DeepFaceLab ? reverse-engineered prompt

Build me a desktop app for creating face swap videos from your own photos and clips, with a simple workflow that feels approachable even if I am not an expert.

I want to be able to load a source face and a target video, extract faces, train a model, preview the result, and then merge it back into the video with decent quality. It should also support things like de aging and head replacement, not just basic face swapping. Make the interface clear, with progress feedback, helpful defaults, and a few guided steps so it does not feel overwhelming.

Use the existing Python codebase and make it work well on a modern GPU setup, with options for CUDA and a fallback path if possible. If you need to check current docs or best practices online, go ahead. Keep the output organized so I can run the main app, process media, and manage models without hunting through a bunch of hidden settings.