JayWebtech/autoshorts — reverse-engineered prompt
Reverse engineered prompt
Build me a local first desktop app called AutoShorts that helps turn long podcasts, interviews, or videos into short vertical clip candidates for social media. I want to import a video or audio file, have the app extract audio, transcribe it, analyze the transcript with AI, find the best viral moments, rank them, and suggest catchy hook ideas. It should save everything locally in projects so I can create, open, rename, and delete projects, and keep transcripts, clip candidates, custom names, and render info.
The app should also auto crop landscape video into 9 by 16 portrait clips and export them in a common format. On first launch, guide me through setup with either cloud API keys for transcription and AI, or a more offline option using local models and Whisper. Show clear warnings in the UI when required keys or tools like ffmpeg are missing. Please make it feel polished, simple, and usable on desktop, and look up current docs online if you need to.
Want more depth? Deep Reverse