msrbuilds/voice-studio ? reverse-engineered prompt

Build me a local web app for making AI voices on my own machine.

I want a simple studio where I can switch between a few different open source voice models, type text and generate speech, or use a podcast style editor where different speakers can each have their own voice. It should also let me drop in an audio file, turn it into an editable transcript, export subtitles, and use that transcript to re voice the audio in another voice. If possible, let me translate the transcript too so I can dub audio into another language.

Please make it work offline after setup, with a clean dark and light interface, voice uploads, built in sample voices, caching so repeated generations are fast, and support for both GPU and CPU if available. If you need to look up current docs online for any model setup details, go ahead.