midudev/subvid.app — reverse-engineered prompt
Reverse engineered prompt
Build me a simple browser app for adding subtitles to video or audio, all on the user’s device with no uploads, no backend, and no API keys. I want people to drag in an MP4, MOV, WebM, MKV, MP3, WAV, or OGG file, choose the spoken language or auto detect it, choose the subtitle language, then generate subtitles with AI. After that they should be able to edit the text and timing in a clean timeline editor with live preview, undo and redo, and a few caption style options like font, color, background, outline, and position. Let them translate subtitles too, support more than one subtitle track, and export either an SRT file or an MP4 with captions burned into the video. Make it feel fast in modern browsers, cache the models after first use, and include English and Spanish pages. If a browser lacks the faster video features, just fall back gracefully. You can check current browser docs online if needed.
Want more depth? Deep Reverse