Subtitles & Captions
Karaoke
Word-by-word karaoke captions
Turn one recording of a person talking into a 9:16 short with karaoke captions, each word lighting up as it is spoken, in Vietnamese or English.
Important: this is a skill for terminal AI coding agents (Claude Code, Codex) that runs on your own computer. It is not a prompt for the ChatGPT website. Check the requirements below before buying.
What actually happens
Karaoke captions are the captions of the short that the Fast cut look (Video Editing skill) renders, so you get that whole pipeline. From one recording of one person talking, the agent:
- Transcribes it, then cuts silences, Vietnamese fillers (à, ờ, ừm), stutters and restarted phrases at light, medium or strong level.
- Picks one story window (45-75 seconds by default) and crops to 9:16 around the speaker's head, with captions above the face.
- Burns in the karaoke captions plus a hook title on the first frame, and renders at -14 LUFS.
The result is shorter and reframed: this look does more than add captions.
How to ask
Thêm phụ đề karaoke từng chữ cho video talk.mp4What you need
- Python 3.10 or newer and ffmpeg (GPU optional).
- A Soniox key, or an ElevenLabs Scribe key, for speech-to-text.
- opencv-python-headless and numpy in a small Python environment, and Edge or Chrome for the title card image.
Who it is for
People with a recording of one person talking (talking head, vlog, podcast) who want a captioned vertical short.
Not a fit if you want
- Captions with nothing else changed: this pipeline cuts and reframes too.
Can I use the videos commercially?
Videos you render are yours for any commercial use. You may not resell, redistribute or share the skill files.
Upgrade path
Start with one look, step up when you need more
More







