You record a live, a webinar or a long video, and nobody has hours to find the best 60 seconds and crop them vertically. Tools that do it charge $20 to $300 a month, according to the README itself. This one takes a YouTube link or your own file, transcribes it with Whisper and asks an AI for the moments with the strongest hook. It hands back 9:16 vertical clips with a score, a hook line and the reason each was picked.
Who this is NOT for
It is not for you if you expect a free Opus Clip with a button: there is no interface, you run it from the terminal with Python 3.10. It also does not fit if your reels need animated captions, because the README never mentions them; you would have to add them separately. And if your video has no face on screen, like a screen share or a product on a table, keep in mind that the local crop tracks faces, and the README does not say what it does when there are none. If you have no long videos recorded yet, there is nothing to clip: start by recording.
What truly sets it apart is not the price, it is that the judgment is visible. Every clip comes with its score and its reason, and the list of virality signals is editable in `highlights.py`. You stop trusting a black box and can teach it what a good moment means for your business. Watch the 'offline' claim: even in local mode, your video's full transcript travels to OpenAI or Gemini. If you record client meetings, that matters. Watch the README too: it pushes toward MuAPI's paid API again and again, and its `git clone` command points to `SamurAIGPT/...`, not `Anil-matcha/...`, so check which repo you are cloning. The project is active (last push on September 17, 2026). The official page it links to does not document the tool: it is a generic article listing other apps.
Your first step
Pick a long video you already recorded and run python main.py "/path/to/your-video.mp4" --mode local --num-clips 3 --language es. First you need requirements-local.txt installed, OPENAI_API_KEY in your .env and ffmpeg on your PATH. Review the 3 clips in ./output/ and read the reason it gave for each one.
The AI YouTube Shorts Generator implementation guide
The project belongs to Anil Chandra Naidu Matcha (Anil-matcha) and it is free: the link beside asks for nothing. The guide to set it up without getting lost is ours: step by step, common mistakes and a checklist. Leave your email and we will send it.
No spam. Unsubscribe anytime.