kaixxx/noScribe ? reverse-engineered prompt

Build me a desktop app for transcribing interview audio locally on my computer.

It should take an audio file, run AI transcription, identify different speakers, and then show me a clean transcript I can review and edit in a simple editor. I want it to feel like a real app, not a command line tool, with a nice main window, clear progress while it works, and an easy way to correct mistakes after the transcript is created.

Please make it work completely offline after setup, since the point is to keep interviews private. It should support multiple languages and be useful for qualitative research and journalism. I’d like it to run on Windows, Mac, and Linux if possible, and use current Whisper and speaker identification models under the hood. If you need to look up current docs or best practices online, go ahead. Keep the interface polished and straightforward, with the main focus on high quality transcription and an editing workflow that makes review fast.