scimotion-video

🎬 SciMotion Video Skill

An OpenClaw agent skill for creating animated explainer videos entirely from code.

No image generation APIs — all visuals are SVG + CSS + GSAP, rendered via Puppeteer to MP4.

What It Does

Script → Azure TTS → Audio + Timestamps → Animation HTML (GSAP) → Puppeteer Render → MP4

1. TTS Generation — Azure Speech Service with configurable voice/style
2. Animation HTML — Self-contained 1920×1080 file with GSAP master timeline
3. Video Rendering — Frame-accurate Puppeteer screenshots → ffmpeg → MP4

Install

# Download the .skill file and install
openclaw skills install scimotion-video.skill

Requirements

• Node.js 18+
• puppeteer (npm)
• ffmpeg + ffprobe
• Azure Speech Service API key (for TTS)

Usage

Tell your agent:

"Make me an explainer video about the sunk cost fallacy"

The skill guides the agent through the full pipeline: script → TTS → animation → render.

File Structure

scimotion-video/
├── SKILL.md                           # Main skill instructions
├── scripts/
│   ├── generate-tts.js               # Azure TTS generator
│   └── render-video.js               # Puppeteer → MP4 renderer
└── references/
    ├── animation-patterns.md          # Reusable GSAP/SVG patterns
    └── script-format.md              # Script JSON schema

Credits

Built by Fogsight 雾象 — AI-powered science video engine.