abus-aikorea/voice-pro ? reverse-engineered prompt

Build me a user friendly web app where I paste a YouTube link and it downloads the video, pulls out the clean voice by removing background music, transcribes the speech accurately with timestamps, translates it to any language I pick, and then dubs everything back using natural sounding voices. Let me upload a short voice sample to clone that exact tone for the new audio so it feels like the same speaker talking in another language. Add simple controls for subtitles, different TTS styles, and batch processing for longer content like podcasts or audiobooks. Make the interface clean with tabs so creators can go step by step or run the full pipeline at once. It should work great on Windows with a GPU and support tons of languages. Focus on speed and quality so I can turn foreign videos into my own dubbed versions quickly.