Recreate an animation in After Effects with AI
You can recreate an animation in After Effects with AI: show the reference to ChatGPT, Claude or any MCP client connected to Prism, and it is rebuilt in your open comp as editable shape layers, text layers and keyframes.
The reference can be an AI-generated clip, a Dribbble shot, an ad, a GIF or a still. Prism never watches it; your AI client does. The client describes each shot as layers, timing and ease, and Prism builds that description in the open comp. Once it is layers, a rebrand is a sentence.
What to give your AI
| Reference | Which client reads it |
|---|---|
| A video, or an AI-generated clip | Gemini and ChatGPT take the video file; Claude takes frames you paste |
| A GIF | any client that sees images, from the frames that show the most |
| A Dribbble shot or an ad | paste the still or a few frames; the .aep is not needed |
| A still | any client that sees images |
What you say, and what lands
| You type | What appears in your comp |
|---|---|
| "Rebuild this in the active comp: three chevron arrows on a circle, 120° apart, made from an ellipse stroke with trim paths and a 3-point polystar. They slide in from off-screen over 0.6 s with a soft overshoot, hold 0.4 s, spin one full turn together, then fill changes from #FF7A1A to white. Transparent background, 1080×1080, 30 fps, everything as native shape layers with editable keyframes." | Three shape layers with Trim Paths and a polystar, position keyframes with overshoot, one shared rotation, a fill colour keyframe |
| "Here are six frames from the AI-generated clip. Rebuild shot one at 1920×1080: the mark draws on, the wordmark fades up beneath it, then both drift 40 px up as the tagline lands." | A shape layer with a Trim Paths write-on, two text layers with opacity and position keyframes, one shared drift |
| "Match the ease: fast out, long settle, like a spring that barely bounces." | The easing rewritten on every keyframe |
Test it on your own project. A Free account is enough to start: Sign up, then connect your AI and ask for this.
How to do it
Rebrand and repurpose
| You want | You say | What changes |
|---|---|---|
| Your colours | "Swap every #FF3D05 fill and stroke to #0A66FF, except the Logo layers" | every matching fill and stroke, see change a colour |
| Your logo | "Replace the logo layer with ~/brand/logo.svg, keep its position and keyframes" | the SVG lands as shape layers in the old layer's place |
| Your type | "Change every text layer to Inter-Bold" | every text layer, see change every font |
| Every format | "Make the 9:16 and 1:1 versions" | new comps with the layout rebuilt, see every format |
What it needs
| After Effects | 2024, 2025 or 2026, macOS or Windows |
| AI client | any MCP client: ChatGPT, Claude, Claude Code, Cursor, Codex and others |
| Plan | a Free account is enough; this uses MCP actions, not the Prism AI balance |
What it does not do
Prism does not watch the video, measure the reference's exact values, or recover its .aep. It copies no footage: the result is layers, not pixels, and the taste stays yours.
Common questions
Whatever your client can see. Gemini and ChatGPT read a video file; Claude reads frames you paste; every client reads a still or a description.
Yes. Every move is keyframes on real properties, with easing you can drag in the graph editor.
Yes. Describe the shot or paste a still of it; the layers and timing are rebuilt from that, not from a project file.
Yes. Cards, icons and marks are shape layers, words are live text layers; the reference never enters the project.
Yes. Once it is layers, a colour swap, a logo replacement, a font change or a 9:16 version is one sentence each.
Related
Animatic
Animatic in After Effects with AI: import your boards one layer each, transcribe the voiceover, and time every shot to its sentence, with shot numbers and a burn-in.
Talking head
Talking head video in After Effects with AI: transcribe the take, then add word-timed captions, icon cards on the keywords, jump cuts and sound as editable layers.