isaacrlevin/TranscriptGenerator — reverse-engineered prompt

Reverse engineered prompt

GitHub

Build me a simple cross platform desktop app that turns video files into text transcripts completely offline. I want it to work on Windows and macOS, let me pick common video formats like MP4, MOV, AVI, MKV, WebM, and similar, then extract the audio locally and run Whisper on my machine.

The app should be friendly for non technical users. On first launch it should check whether FFmpeg is installed and clearly explain what to do if it’s missing. Let me choose and download a Whisper model size, from fast and small to slower and more accurate. Then I should be able to browse for a video, choose whether timestamps are included, generate the transcript with progress shown, and save the result with a normal save dialog that defaults to Downloads.

Please make privacy clear in the UI, videos and transcripts should never be uploaded anywhere. Also make the saved transcript path clickable so I can open it easily. Use the current .NET MAUI and Whisper.NET docs if needed.

Want more depth? Deep Reverse