Claude Code Turns Any Topic Into Animated Explainer Videos Instantly
A Claude Code skill reverse engineers viral Douyin knowledge shorts, generating storyboards and rendering finished videos frame by frame with Canvas and Playwright.
Takeaways
- vibe-knowledge-film is a Claude Code skill that generates Douyin style knowledge explainer shorts.
- Two preset styles: Nocturne for hard science topics, Fable for behavioral and psychological concepts.
- Renders frame by frame with Canvas, Playwright, and ffmpeg, not a video diffusion model.
- Agent handles script, storyboard, subtitle track, sound effects, and a 30 second preview cut.
- MIT licensed engine and scripts, with separate licensing for reference analyses from third party videos.
- Requires local Chrome, ffmpeg, Playwright, and installs into
~/.claude/skills/.
Developer HeyChengC has released vibe-knowledge-film, a Claude Code skill that generates short, animated explainers from a topic or finished script. The project recreates a format popular on Chinese short-video platforms, where dense lessons on subjects such as protein folding and behavioral psychology use kinetic typography, diagrams, narration, and tightly synchronized motion graphics. The repository has 846 GitHub stars.
The project packages the creative workflow as code. Claude selects a visual system, writes subtitles, plans shots, scaffolds the rendering project, and produces a preview before rendering the full video. Developers get editable, repeatable output without training or running a video-generation model.
Two visual systems, two kinds of story
HeyChengC derived two presets from analyses of popular examples. Each preset constrains typography, color, composition, transitions, and pacing so the generated scenes share a coherent visual language.
| Preset | Flag | Best suited to |
|---|---|---|
| Nocturne | C |
Abstract ideas, technical subjects, and hard science |
| Fable | B |
Behavioral and psychological concepts that benefit from narrative scenes |
The --world option selects the preset when creating a film. These constraints keep the renderer focused on two established aesthetics instead of attempting arbitrary visual styles.
Canvas frames replace generated footage
JavaScript and Canvas draw every frame programmatically. Playwright controls a headless browser and captures the rendered sequence, then ffmpeg combines the frames, narration, generated sound effects, and user-supplied background music into the final video.