wendy7756/AI-Video-Transcriber — reverse-engineered prompt
Reverse engineered prompt
Build me a simple local web app where I can paste a video or podcast link, or upload a file, and get an AI cleaned transcript plus a good summary.
It should work with common sources like YouTube, TikTok, Bilibili, Apple Podcasts, SoundCloud, and similar sites, and also accept local audio, video, or plain text files. If a platform already has subtitles, use those first so it feels fast. If not, fall back to speech to text. After transcription, clean up the wording, fix obvious mistakes, and format it into readable paragraphs.
I want to choose the summary language, and if that language is different from the source, automatically create a translation too. Please include a simple AI settings area in the UI where I can enter an OpenAI compatible base URL and API key, fetch available models, and pick one. Show clear progress while it runs, make it mobile friendly, and let me download the transcript, translation, and summary as Markdown files.
Please make it easy to run locally, and look up current docs online if needed.
Want more depth? Deep Reverse