Creating a Vox‑Style Paper‑Cut Video with Codex: A Zero‑Skill Walkthrough

Hello, I’m Yingliu. AI‑generated Vox‑style videos have become popular on X, so I tried the open‑source project paper-collage-ad-codex together with Codex.

I aimed to create a short paper‑cut video titled “The Development of Chinese Art History”. After downloading the Skill, I supplied two reference X videos and the prompt to make the video.

https://x.com/xiaoerzhan/status/2076140434653585568/video/1, https://x.com/xiaoerzhan/status/2076518889060782460/video/1 参考这两个视频，帮我制作一条，中国美术史的发展。

Codex first generated a storyboard and narration, which I reviewed. Once satisfied, I let Codex continue.

It then produced the first‑frame image and, in a single step, rendered the full video.

The Skill is not truly one‑step; I had to engage in many dialogue rounds and adjust prompts extensively. Codex also ran slowly, and the final video quality was only average.

If you omit reference links and give only a simple sentence, the output is poor. When the multi‑round result is still unsatisfactory, harsher prompting can sometimes improve the result.

Finally, I watched a video made by the original author of the Skill. AI can indeed automate storyboard, narration, and rendering, lowering the entry barrier for non‑editors, but it is still far from a true “one‑click” solution. It’s worth experimenting with, but don’t overhype its capabilities.

Hope this sharing helps you!