A demo for each of the 15 repos from the post, grouped the same way. Where a repo ships its own demo, that clip is shown. Where it doesn't, the tool was run or applied here, today, and the result is shown instead.
Give it a product and it writes the script, adds the voice-over, moves the camera and grades the colour through to a finished launch film.
Clips are the v2.0.0 release files the author published.
github.com/AbubakrChan/product-launch-motion13 skills for launch videos, animated charts and short videos. The author recommends running it on Opus 5.5.
github.com/charlie947/motion-graphics-skillsCuts picture and sound on the beat of the music. This is the author's own launch video, made on Opus 5.5 in 5 rounds.
The repo only ships a small preview GIF, so the picture is low resolution and silent.
github.com/howseen-ai/claude-motion-designGive it raw footage and it cuts the pauses and flubs, tightens the story, then adds captions, B-roll and motion graphics.
The reel is the creator's own talking-head recording after an edit pass with the kit.
github.com/nateherkai/hyperframes-student-kitDrops an animation in exactly where you are talking about something, and can separate a person from the background so graphics sit behind them.
github.com/Barty-Bart/motion-graphicsShort explainers in 7 styles such as cut paper, riso print and pixel art, plus a custom look of your own. It shows you the storyboard and waits for approval before it renders.
The repo ships these two images and no video.
github.com/cth9191/animateGives the AI taste rules, makes it review its own rendered frames for problems, and lets it make its own sound effects.
The repo's previews are small GIFs, so these are silent.
github.com/whaleyxbt/claude-motionFixes the stiff, robotic movement AI-made Remotion videos tend to have, using eased and spring motion.
github.com/haidrrrry/claude-remotion-skillThe animation library itself: buttons, cards and page transitions on a site. Here it drives a staggered card entrance, a spring toggle and a page slide.
github.com/motiondivision/motionA Claude Code skill: describe the animation in one sentence and it writes the Motion code. This stage uses the skill's own "hero fade-up" and "magnetic button" recipes, with its exact timing values.
github.com/199-biotechnologies/motion-dev-animations-skillHeyGen's open-source renderer: an AI writes a web page with HTML, CSS and GSAP, and HyperFrames records it frame by frame into an MP4.
Installed HyperFrames 0.8.153, wrote one HTML page (starfield, planet, moon impact, counter, call to action) and ran hyperframes render. Zero credits.
Remotion itself: videos written as React components and rendered to MP4. The same engine your Shorts pipeline already uses.
Installed Remotion, wrote one composition (planets pop in on springs, diameters count up from NASA figures) and ran remotion render.
// Rules from remotion-dev/skills → remotion-markup, as applied in demo 12
const {fps} = useVideoConfig(); // fps from useVideoConfig(), never hard-coded
const frame = useCurrentFrame(); // every motion driven by the frame number
const pop = spring({frame, fps, // spring() instead of CSS transitions,
config: {damping: 12, stiffness: 120}}); // which don't render in Remotion
const count = interpolate(frame, [0, 1.2 * fps], [0, km], {
extrapolateRight: 'clamp',
easing: Easing.out(Easing.cubic), // Easing.* to shape the timing
});
<Sequence from={(0.6 + i * 0.8) * fps} // timing in seconds × fps
premountFor={fps}> // one second of premount on timed items
<Planet {...p} />
</Sequence>
Remotion's official manual for AI agents. It is a set of rules, not a renderer, so its demo is the code above: the Remotion clip in 12 was written to these rules.
The skill also covers voice-over, captions, SFX, transitions, maps and 3D.
github.com/remotion-dev/skillsGreenSock's official skills. This sequence follows the timeline skill's rules: one timeline, shared defaults, labels and position parameters instead of chained delays.
const tl = gsap.timeline({
defaults: {duration: .5, ease: "power2.out"}});
tl.addLabel("intro", 0)
.from(words, {y: 30, autoAlpha: 0, stagger: .08}, "intro")
.addLabel("bars", "-=0.1")
.from(bars, {scaleX: 0, stagger: .15}, "bars")
.from(nums, {textContent: 0, snap: {textContent: 1}}, "bars");
github.com/greensock/gsap-skills
LottieFiles' Disney-principles skill for AI. It works with any tool. Each lane plays the same entrance using the duration, easing and overshoot the skill assigns to one of its four "motion personalities".
Pick one personality per project and keep it, the skill says. Compare how differently the same box feels.
github.com/LottieFiles/motion-design-skill