Repolode
Submit
alchaincyf

alchaincyf/huashu-art-motion

An agent skill that lets coding agents turn 35+ art styles and explainer-video grammars into code-drawn animation.

Open-source alternative to:After EffectsAdobe Animate

View on GitHub
Stars
4,220
Forks
457
Open issues
7
Last commit
2 days ago

What is it?

huashu-art-motion is an agent skill for producing animation entirely in code, so paintings and illustrations actually move. It includes 36 art-style recipes (such as Van Gogh, Egyptian murals, Bauhaus, and 8-bit pixel art), each with a matching Canvas scene, along with 8 explainer-video grammars like Kurzgesagt, Vox, whiteboard, and 3Blue1Brown style. Clips are driven by a JSON spec and rendered frame-accurately in landscape, portrait, or transparent formats.

The project also provides a long-scroll skeleton in which a character walks through a sequence of paintings, switching art style at each boundary, plus a breakdown script that maps a reference animation into cuts, beat grids and motion heatmaps. A QA script measures stability, per-frame cost, motion and smoothness before delivery. Character imagery is intended to come from an external image model, with the code handling compositing and材质.

The skill is written in Chinese and requires uv, ffmpeg and Playwright Chromium. Code and docs are MIT licensed, while bundled fonts and certain character assets carry their own terms.

Key features

  • 01Provides 36 art-style recipes, with 35 matching Canvas scenes.
  • 02Includes 8 explainer-video grammars with parameterized clips driven by JSON specs.
  • 03Renders frame-accurately in landscape, portrait, or transparent output.
  • 04Offers a long-scroll skeleton for a character moving through styled worlds.
  • 05Breaks down reference animations into cuts, beat grids and motion heatmaps.
  • 06Runs a QA script that measures stability, cost, motion and smoothness.

Use cases

  • Creating stylized animated explainers for YouTube-style narration videos.
  • Recreating the motion of a reference animation through code rather than keyframing.
  • Producing art-history or painting-themed motion pieces with an agent.
  • Prototyping motion graphics in code with frame-exact timing.

Topics

#animation#generative-art#canvas#agent-skill#video-rendering#javascript#motion-graphics

FAQ

What is huashu-art-motion?
An agent skill that lets coding agents turn 35+ art styles and explainer-video grammars into code-drawn animation. huashu-art-motion is an agent skill for producing animation entirely in code, so paintings and illustrations actually move. It includes 36 art-style recipes (such as Van Gogh, Egyptian murals, Bauhaus, and 8-bit pixel art), each with a matching Canvas scene, along with 8 explainer-video grammars like Kurzgesagt, Vox, whiteboard, and 3Blue1Brown style. Clips are driven by a JSON spec and rendered frame-accurately in landscape, portrait, or transparent formats.
Is huashu-art-motion free to use? What license does it have?
huashu-art-motion is open source under the MIT license. Check the license terms before using it in commercial products.
Is huashu-art-motion still maintained?
Actively maintained. The last commit to huashu-art-motion was on Oct 16, 2026.
What can I use huashu-art-motion for?
Creating stylized animated explainers for YouTube-style narration videos. Recreating the motion of a reference animation through code rather than keyframing. Producing art-history or painting-themed motion pieces with an agent. Prototyping motion graphics in code with frame-exact timing.
Where can I get huashu-art-motion?
The source code and installation instructions are on GitHub: https://github.com/alchaincyf/huashu-art-motion

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects