A Claude Code skill that turns a short theme into a beat-synced motion graphics video rendered to MP4 with HyperFrames.
siyuanfeng636-cpu/agentic-motion-graphics
A Claude Code skill and Python pipeline that generates motion graphics videos from HTML, GSAP animations, TTS narration, and Whisper word-level timing.
- Stars
- 9
- Forks
- 0
- Open issues
- 0
- Last commit
- 27 days ago
What is it?
Agentic Motion Graphics is a skill for Claude Code and similar AI agents that produces videos as code rather than as generated footage. The agent writes a scene script, builds each frame with HTML, CSS, SVG and GSAP timelines, and renders the result to MP4 with the HyperFrames CLI in headless Chromium.
The pipeline handles audio as the timing reference. Voiceover is generated through OpenAI TTS or ElevenLabs, trimmed with ffmpeg, and aligned to individual words with whisper.cpp or faster-whisper so animations can be bound to exact timestamps. Background music and UI sound effects are synthesized in Python with numpy and scipy on a 120 BPM grid.
The repository includes scripts for TTS, timing alignment and sound synthesis, a boilerplate project for HyperFrames, and prompt templates covering product launches, explainer videos, vertical short-form clips, client proposals and slide animation.
Key features
- 01Builds each frame from HTML, CSS, SVG and GSAP timelines instead of generative video.
- 02Aligns narration to word-level timestamps with whisper.cpp or faster-whisper.
- 03Generates voiceover with OpenAI TTS or ElevenLabs and trims silences with ffmpeg.
- 04Synthesizes 120 BPM music beds and UI sound effects in Python.
- 05Runs lint, snapshot and headless render steps through the HyperFrames CLI.
- 06Packaged as a skill directory that Claude Code can discover automatically.
Use cases
- Producing product launch or feature announcement videos from a website or changelog.
- Turning a long technical article or paper into an animated explainer with narration.
- Converting long videos into captioned vertical clips with word-by-word subtitles.
- Animating static HTML or Markdown slides for talks.
Quick start
- 1
Install FFmpeg and Node.js 22+ (macOS example with Homebrew)
brew install ffmpeg node - 2
Generate the voiceover audio with the TTS script
python3 scripts/tts_engine.py --text "Opus 5.5 solved motion graphics with code." --output audio/voiceover.mp3 - 3
From templates/boilerplate, render the project to MP4 with HyperFrames
npx hyperframes render --output dist/output.mp4 --fps 30
Things to know
- The README's setup examples are macOS-focused (Homebrew), so other platforms need their own setup.
- The TTS step needs an OpenAI or ElevenLabs API key, which is a paid service.
- The pipeline has many dependencies, including Node.js 22+, FFmpeg, Whisper and Python audio tools.
Topics
FAQ
- What is agentic-motion-graphics?
- A Claude Code skill and Python pipeline that generates motion graphics videos from HTML, GSAP animations, TTS narration, and Whisper word-level timing. Agentic Motion Graphics is a skill for Claude Code and similar AI agents that produces videos as code rather than as generated footage. The agent writes a scene script, builds each frame with HTML, CSS, SVG and GSAP timelines, and renders the result to MP4 with the HyperFrames CLI in headless Chromium.
- Is agentic-motion-graphics free to use? What license does it have?
- agentic-motion-graphics is open source under the MIT license. Check the license terms before using it in commercial products.
- Is agentic-motion-graphics still maintained?
- Actively maintained. The last commit to agentic-motion-graphics was on Sep 25, 2026.
- What can I use agentic-motion-graphics for?
- Producing product launch or feature announcement videos from a website or changelog. Turning a long technical article or paper into an animated explainer with narration. Converting long videos into captioned vertical clips with word-by-word subtitles. Animating static HTML or Markdown slides for talks.
- How do I install agentic-motion-graphics?
- Install FFmpeg and Node.js 22+ (macOS example with Homebrew): brew install ffmpeg node Generate the voiceover audio with the TTS script: python3 scripts/tts_engine.py --text "Opus 5.5 solved motion graphics with code." --output audio/voiceover.mp3 From templates/boilerplate, render the project to MP4 with HyperFrames: npx hyperframes render --output dist/output.mp4 --fps 30
- Where can I get agentic-motion-graphics?
- The source code and installation instructions are on GitHub: https://github.com/siyuanfeng636-cpu/agentic-motion-graphics
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
A curated collection of 18 full prompts and 54 public cases showing how Claude Opus 5.5 generates videos by writing code.
An Agent Skill that turns a topic into a narrated explainer video with voiceover, hard subtitles and cover images, rendering HTML frames locally to MP4.
Turns a written script into a narrated explainer video or product demo, with a local voice, word timing and canvas rendering, driven by a coding agent.
A set of Claude Code skills that make videos by generating every frame and music track in code, with one skill per visual style.
A Claude Code / Codex agent skill that turns a product description into a commercial-grade launch film, built from code with a composed score and point-and-click editing.