Motion Graphics
Creative skill, available on Zeplik
Motion Graphics is a ready-to-run creative skill on Zeplik. Animated titles, kinetic type, logo stings, lower thirds, charts, counters, diagram reveals and loops -- built as a self-playing HTML artifact. Ask in plain language and Zeplik applies the skill's method for you inside the conversation, on whichever AI model you prefer. It returns a structured code you can keep and reuse: Single-file HTML artifact -- a self-playing animation that runs live in the artifact pane and downloads as a standalone .html.
The Motion Graphics skill loads automatically when your request matches it, or you can invoke it directly by typing /motion-graphics in any chat. It works with attachments, connectors, and any model that supports the task, so you get the same expert method every time without setting anything up.
What the Motion Graphics skill can do
- Build a self-playing animation as one HTML file: title cards, kinetic type, logo stings, lower thirds
- Animate real data -- bars growing from the baseline, a number counting up, a line drawing itself
- Drive every element from one clock and one beat table, so the timing stays re-editable in place
- Choose the easing, stagger and hold times that separate designed motion from a CSS demo
Try these prompts on Zeplik
Pick a prompt to open it in the Zeplik app. If you are not signed in yet, your prompt is waiting for you the moment you do.
How the Motion Graphics skill works
/motion-graphics
Motion graphics is designed motion: type, shape, colour and data moving on purpose. That makes it the opposite of generated footage in every way that matters — it has to be frame-exact, the words have to be spelled right, the brand colour has to be that hex, and it usually has to loop. A diffusion model is the wrong instrument for all four.
So the default deliverable here is a single-file html artifact that plays
itself: canvas, SVG and CSS driven from one clock. It runs live in the artifact
pane, the user can replay it and download the .html, it costs nothing, it has no
length limit, and every frame is exactly what you specified.
Route before you build
Designed motion → build it here. Titles, kinetic type, logo stings, lower
thirds, animated charts and counters, diagram reveals, hero loops, transition
wipes. Recipes: references/recipes.md. Runtime contract and the clock pattern:
references/artifact-runtime.md.
Photographed or photoreal motion → video-generation. A drone shot, a coffee
pour, smoke, an abstract organic texture. Anything whose appeal is that it looks
filmed. That path is paid per second; do not reach for it to animate a word.
A project they render to a real file themselves → remotion (React) or
motion-canvas (TypeScript). This is the honest answer whenever the deliverable
must be an mp4 at broadcast quality: they get code, they run the render. For a
mathematical animation in the 3Blue1Brown idiom, manim.
A narrated, teaching sequence → explainer-video. If the point is to explain
something rather than to make something move, that skill owns the structure,
narration and assembly.
The boundary, stated plainly
Nothing in this product encodes a rendered page into an mp4. The render
sandbox has ffmpeg and a font in it, no browser. So an artifact built here plays
in the pane and downloads as .html; it does not come out the other side as a
file to upload to a social platform.
Say that in one sentence when it matters, and offer the two real routes: screen-
record the artifact while it plays, or go the remotion / motion-canvas way and
render properly. Never paper over it by generating footage that vaguely resembles
the animation they asked for — the type will be wrong and they will be billed for
it.
What the video timeline gives you, so you do not over-promise
If there is already footage in the conversation, edit_video can put static
text overlays on it — nine-point grid placement, a size, a colour, an optional
box, up to 8 per clip and 200 characters each — plus transitions between shots
from a fixed set. That is titling, and for a subtitle or a name caption over video
it is the right tool.
What the timeline does not have: keyframes, transforms, per-clip animation,
pan or zoom on a still. Text there appears and disappears; it does not fly in,
type on, or count up. So a title card is built here as an artifact; a name
caption over footage goes on the timeline. If the user wants animated lettering
burned into their video, that is the remotion route — say so rather than
delivering a static overlay and calling it a motion title.
Craft that separates this from a CSS demo
- One clock, and timing in one table. A pile of
@keyframeswith staggeredanimation-delaycannot be replayed, paused, or re-timed, and a late webfont desynchronises all of it. Drive everything from a singlerequestAnimationFrameloop against a declarative list of beats.references/artifact-runtime.mdhas the pattern. - Easing is the whole difference. Linear motion is what makes an animation look cheap, instantly. Nothing moves linearly: ease out for things arriving, ease in for things leaving, and a slight overshoot on anything that should feel physical. Never ease a fade the way you ease a move.
- Offset, do not synchronise. Six words arriving together is a block moving; the same six words 60–80ms apart is kinetic type. Stagger is the effect.
- Fast in, slow to settle. Entrances around 300–500ms, holds of at least 1.2s so the eye can actually read, exits faster than entrances.
- Type is the design. Weight, tracking, scale contrast and one accent colour carry a title card. Pull a real webfont from a CDN and name a fallback.
- Restraint. Animate one property per element where you can. Three elements moving deliberately beat nine moving at once, and the nine will also drop frames.
- Loops must actually loop. If it is a hero loop, the last frame has to equal the first — build the timeline modulo its own duration rather than ending and restarting.
Workflow
- Get the content exact before animating it. The words, the numbers, the brand hex, the logo shape. Motion cannot rescue wrong copy, and re-animating after a copy fix is the commonest waste here.
- Name the format. Aspect ratio and duration: a 16:9 title card, a 1:1 loop, a 9:16 opener. It changes the layout, not just the canvas.
- Write the beat table first — element, in-time, duration, easing. Show it if the piece is complex; it is the thing the user actually has opinions about.
- Build one fenced
htmlartifact, self-contained, responsive, with a replay control. - Iterate one property at a time. "Slower" and "punchier" are timing and easing changes in the beat table, not rewrites.
Guardrails
- Never fake it. No invented video link, no screenshot passed off as an animation, no "the animation is attached". The artifact is written out in full or it does not exist.
- Use the user's own marks. Their logo, their colours, their type. Do not reconstruct another company's logo or a copyrighted character, and do not pull in copyrighted music.
- Carry visuals with inline SVG, CSS and type. Never invent an image URL or host —
a made-up
srcpaints a broken block. - No fixed width past ~360px; the artifact must not overflow at 390px.
- Animate
transformandopacity, nothing else.width,height,top,left,marginandbox-shadowforce layout or paint on every frame. A bar that grows isscaleYon atransform-origin: bottomrect with the label counter-scaled, never an animatedheight; a wipe is atranslateor aclip-path, never an animatedwidth. - Honour
prefers-reduced-motion, in the clock and not only in CSS. Read it once withmatchMedia("(prefers-reduced-motion: reduce)")and, when it is set, paint the FINAL composed frame and never start the loop. A CSS media query cannot save you here: the loop writes inline styles every frame, and an inline style outranks any rule you wrote —*{animation-duration:.01s}does not stop arequestAnimationFrametimeline at all. Freezing on frame zero is the other failure; frame zero is usuallyopacity: 0, which shows the viewer nothing. - Audible audio does not autoplay in the artifact frame, so never build a piece whose timing depends on sound starting by itself.
Usage
/motion-graphics $ARGUMENTS
How to use the Motion Graphics skill
Sign in to Zeplik
Create a free Zeplik account or sign in. New accounts start with free credits, so you can try the Motion Graphics skill right away.
Describe your creative task
Ask in plain language, or type /motion-graphics to invoke the skill directly. Zeplik recognizes the Motion Graphics skill and applies its method.
Review and refine the result
Zeplik returns a structured code you can edit, download, and reuse. Ask follow-ups to refine it.
Source and credit
- Author
- zeplik
- License
- MIT
Original Zeplik skill. Built and maintained by the Zeplik team.
Frequently asked questions
- What is the Motion Graphics skill?
- Motion Graphics is a ready-to-run creative skill on Zeplik. Animated titles, kinetic type, logo stings, lower thirds, charts, counters, diagram reveals and loops -- built as a self-playing HTML artifact. Ask in plain language and Zeplik applies the skill's method for you inside the conversation, on whichever AI model you prefer. It returns a structured code you can keep and reuse: Single-file HTML artifact -- a self-playing animation that runs live in the artifact pane and downloads as a standalone .html.
- How do I use Motion Graphics on Zeplik?
- Sign in to Zeplik and ask in plain language, or type /motion-graphics in any chat to invoke it directly. The skill applies its method and returns a result you can refine in the same conversation.
- Which AI model does the Motion Graphics skill use?
- Any model you choose. Zeplik works across every model in one chat, so the Motion Graphics skill runs on your preferred model for the task.
- Where does the Motion Graphics skill come from?
- The Motion Graphics skill is an original Zeplik skill, maintained by the Zeplik team.
- How much does the Motion Graphics skill cost?
- Using the skill is free to start. You only spend Zeplik credits when the assistant runs, and new accounts begin with free credits.
Related creative skills
- 3D Web ExperiencesUse when building 3D on the web: Three.js, React Three Fiber, WebGL, Spline, scroll-driven 3D scenes. Not for 2D browser games (use develop-web-game) or non-3D artifacts (use web-artifacts-builder).
- AI Image GenerationWrite high-quality prompts for Black Forest Labs FLUX image models - text-to-image, edit instructions, structured JSON scene prompts, exact hex colors, text rendering and multi-reference composition.
- Algorithmic Art StudioUse when the user asks for generative or algorithmic art made with code - p5.js sketches, flow fields, particle systems, seeded randomness with interactive parameter exploration.
- Explainer VideoExplain something as a video -- a narrated sequence cut on the timeline, or a self-playing drawn-on board. Not a static hand-drawn sketch (use excalidraw).
- Game Development StudioUse when designing or building games in engines like Unity, Godot, or Unreal -- mechanics, GDScript, ECS/DOTS, game systems, performance. Not for browser games built in chat (use develop-web-game).
- Image EnhancerUse when the user uploads an image to improve -- sharpness, clarity, lighting, composition, artifacts -- via analysis plus a regeneration prompt. Not for new images from scratch (use imagegen).
More on Zeplik
Try Motion Graphics on Zeplik
Every model, one chat. Bring the Motion Graphics skill into your next conversation and let the assistant do the work.