Programmatic video
Skill itallstartedwithaidea/agent-skills/skills/media-creative/programmatic-video
Programmatic Video enables React-based video rendering with Remotion, where every frame is a React component and every animation is expressed in code.From its SKILL.md
npx -y skills add itallstartedwithaidea/agent-skills --skill programmatic-videoAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
SKILL.md
6.4 KB, ~1.5k tokens by cl100k_base, as published. Nobody here has run it
Programmatic Video
Part of Agent Skills™ by googleadsagent.ai™
Description
Programmatic Video enables React-based video rendering with Remotion, where every frame is a React component and every animation is expressed in code. The agent builds video compositions with sequences, transitions, spring animations, and data-driven content, producing MP4/WebM output rendered in headless Chromium without any video editing software.
Traditional video editing is manual, non-reproducible, and impossible to template. Programmatic video treats video as code: compositions are versioned in git, parameterized for batch rendering, and tested with snapshot assertions. A single template can generate thousands of personalized videos by changing the input data—product demos, social media clips, onboarding sequences, or marketing videos, each customized per recipient.
Remotion's architecture maps React's component model to video: each component receives a frame number and the total durationInFrames, using these to calculate animations, transitions, and timing. The agent composes scenes from reusable components, orchestrates their timing with <Sequence> wrappers, and applies spring-based physics animations for natural motion.
Use When
- Creating videos from code without video editing software
- Generating personalized or data-driven video content at scale
- Building animated explainers, product demos, or social media clips
- Automating video production in CI/CD pipelines
- The user requests "programmatic video", "Remotion", or "video from code"
- Producing video assets with consistent branding across hundreds of variants
How It Works
graph TD
A[Video Specification] --> B[Define Composition: FPS + Duration]
B --> C[Build Scene Components]
C --> D[Arrange with Sequences]
D --> E[Add Spring Animations]
E --> F[Inject Dynamic Data]
F --> G[Preview in Browser]
G --> H{Approved?}
H -->|No| C
H -->|Yes| I[Render: Headless Chromium]
I --> J[Encode: MP4 / WebM]
J --> K[Output Video File]
Each scene is a React component that receives the current frame number. Compositions define the canvas size, frame rate, and total duration. Sequences place scenes at specific time offsets. Rendering captures each frame as a screenshot and encodes them into video.
Implementation
import { AbsoluteFill, Composition, Sequence, useCurrentFrame, useVideoConfig,
interpolate, spring } from "remotion";
const TextReveal: React.FC<{ text: string }> = ({ text }) => {
const frame = useCurrentFrame();
const { fps } = useVideoConfig();
const opacity = interpolate(frame, [0, 15], [0, 1], { extrapolateRight: "clamp" });
const translateY = spring({ frame, fps, config: { damping: 12, stiffness: 200 } });
const y = interpolate(translateY, [0, 1], [40, 0]);
return (
<AbsoluteFill style={{ justifyContent: "center", alignItems: "center" }}>
<h1 style={{
fontSize: 72,
fontWeight: "bold",
color: "white",
opacity,
transform: `translateY(${y}px)`,
}}>
{text}
</h1>
</AbsoluteFill>
);
};
const DataDrivenSlide: React.FC<{ metric: string; value: number }> = ({ metric, value }) => {
const frame = useCurrentFrame();
const displayValue = Math.round(interpolate(frame, [0, 45], [0, value], { extrapolateRight: "clamp" }));
return (
<AbsoluteFill style={{ justifyContent: "center", alignItems: "center", background: "#0a0a0a" }}>
<div style={{ textAlign: "center" }}>
<div style={{ fontSize: 120, fontWeight: "bold", color: "#3b82f6" }}>
{displayValue.toLocaleString()}
</div>
<div style={{ fontSize: 32, color: "#a1a1aa", marginTop: 16 }}>{metric}</div>
</div>
</AbsoluteFill>
);
};
export const MainComposition: React.FC<{ title: string; metrics: Array<{ name: string; value: number }> }> = ({
title, metrics
}) => {
return (
<AbsoluteFill style={{ background: "#0a0a0a" }}>
<Sequence durationInFrames={60}>
<TextReveal text={title} />
</Sequence>
{metrics.map((m, i) => (
<Sequence key={m.name} from={60 + i * 75} durationInFrames={75}>
<DataDrivenSlide metric={m.name} value={m.value} />
</Sequence>
))}
</AbsoluteFill>
);
};
export const RemotionRoot: React.FC = () => (
<Composition
id="MainVideo"
component={MainComposition}
durationInFrames={300}
fps={30}
width={1920}
height={1080}
defaultProps={{
title: "Q1 Performance",
metrics: [
{ name: "Page Views", value: 18247 },
{ name: "Conversions", value: 1432 },
{ name: "Revenue", value: 89500 },
],
}}
/>
);
# Preview in browser
npx remotion studio
# Render to MP4
npx remotion render MainVideo output.mp4
# Batch render with different data
npx remotion render MainVideo --props='{"title":"Q2","metrics":[...]}'
Best Practices
- Use
spring()for natural-feeling animations instead of linear interpolation - Clamp all
interpolate()calls to prevent values from overshooting - Keep compositions at 30fps for standard content, 60fps for smooth motion
- Preload all assets (images, fonts) before rendering to avoid missing frames
- Use
<Sequence>to organize scenes with explicit timing, not frame-based conditionals - Test video output at multiple resolutions (1080p, 720p, square) for platform compatibility
Platform Compatibility
| Platform | Support | Notes |
|---|---|---|
| Cursor | Full | React + Remotion development |
| VS Code | Full | Remotion extension available |
| Windsurf | Full | React-based development |
| Claude Code | Full | Component + config generation |
| Cline | Full | Video pipeline setup |
| aider | Partial | Code generation only |
Related Skills
Keywords
programmatic-video remotion react-video animation data-driven-video batch-rendering composition spring-animation
© 2026 googleadsagent.ai™ | Agent Skills™ | MIT License
What ships with it
Read from the repository
Just SKILL.md. No reference files, no scripts.
Gives 0 of the 12 instructions most video audio skills give in ~1.5k tokens
Counted across 619 of the 725 authors here whose files we hold, read 2026-09-06
- Read product marketing context firstin 13 of 619, across 7 files
- Define the core visual thesis in one sentencein 11 of 619, across 3 files
- Break the concept into 3 to 6 scenesin 11 of 619, across 3 files
- Render the smallest working version firstin 11 of 619, across 3 files
- Start with a low-quality smoke test renderin 11 of 619, across 3 files
- Add captions for accessibility and engagementin 11 of 619, across 5 files
- Write the scene outline before writing codein 11 of 619, across 3 files
- Specify subject, action, camera, style, and moodin 11 of 619, across 5 files
- Decide what each scene provesin 10 of 619, across 2 files
- Export one clean thumbnail framein 10 of 619, across 2 files
- Pick the right tool for the jobin 10 of 619, across 4 files
- Run the test suite before proposing a fixin 8 of 619, across 7 files
Said here and by no other author read
- Use spring for natural-feeling animations
- Clamp all interpolate calls
- Keep compositions at 30fps for standard content
- Preload all assets before rendering
- Test video output at multiple resolutions
Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.