
自媒体视频动效 · 人物后大字接全屏宣言
装上下面任意一个 Skill,再让你的 Agent 照着做一条这样的视频。
作者只公开了部分提示词。
Opus 5.5 is an absolute BEAST in creating videos. I am extremely shocked because my previous attempts with GPT-6 Astra and Grok 4.6 were quite underwhelming, far from production ready video. The leap here is gigantic. This is a ZERO shot video. 【作者补充制作说明】 Why Opus 5.5 is so impressive is because it's is not a video generation model. It uses pure code to generate the sound effects, the music. All of the animations are done with CSS/HTML/JS, it doesn't use any video or image assets. This is extremely useful for creating detailed explainer videos where you need very fine control over the infographics and consistent deterministic output. The framework I used was hyperframes and narration is done by OpenAI text-to-speech.