[Intro] Opus 5.5 video-making prompts
Translated by AIView original
A collection of public cases where Claude Opus 5.5 writes code to make videos, with Korean prompts
Opus 5.5 doesn't output video files directly. It writes code such as HTML Canvas, p5.js, Three.js, Remotion, Manim and Blender to build the scenes. You can capture frames in the browser and stitch them into an MP4 with FFmpeg. For music you can use Web Audio or Python synthesis, and for narration you can use TTS.
This approach fits motion graphics, line-art and ink-wash animation, explainer videos, product launch videos and lyric music videos well. If you need realistic footage of people, there's also a workflow where you call video models like Seedance, Runway or Higgsfield and then edit the generated clips.
This repo collects public cases posted during Opus 5.5's first public week (September 22–25, 2026) on X, Bilibili, Linux.do, GitHub, YouMind and others. The prompts are Korean translations and localizations, and each file notes the author and the original link. Some originals were excerpted or cut off when reposted, so each file says so. The times and costs in the cases are the figures the posters reported and don't guarantee the same results.
Comments 0
No comments yet
Be the first to comment.