Skip to content
Prism
Use cases

Explainer video in After Effects with AI

Make an explainer video in After Effects with AI: paste the script into ChatGPT, Claude or any MCP client connected to Prism, and every sentence becomes a scene of shape layers and text in your comp, timed to the voiceover.

Thousands of explainer templates sell the same thing: somebody else's scenes with your words squeezed in. Here the script is the brief, and every scene is built to its sentence.

From a script to scenes

Paste the script and ask for a scene plan before anything is built: one scene per sentence, what is on screen, what moves, which words appear. Edit that plan in the chat; then each scene is built in the open comp, and the voiceover's word timestamps decide when it starts.

What lands

Script lineScene you agree onWhat Prism builds
"Every week, your team loses hours to reports."a calendar with seven cells filling, a clock beside it drainingshape layers, staggered opacity keyframes, a rotation keyframe on the hand
"Acme pulls the numbers for you."the logo, three cards sliding in beneath itthe logo SVG as shape layers, three cards staggered 120 ms apart
"So you get back to the work."the clock refilling, a person icon settlingthe same clock reversed, one icon precomp with an overshoot

Each scene is a precomp with its in point on its first word. Icons are shape layers, drawn or imported from an SVG you supply; words are live text layers in your font.

What you say, and what lands

You typeWhat appears in your comp
"Here is the script. Propose one scene per sentence, what is on screen and what moves. Do not build yet."A scene plan in the chat, nothing in the comp yet
"Build scene 1 in a new 1920×1080 comp, all #1A1A1A on #F5F1E8."A precomp with the calendar and clock as shape layers, keyframed
"Transcribe the voiceover layer and start each scene on its sentence."A [Transcript] guide layer with a marker per word, each scene's in point on its first word
"Scene 5 needs a plate: a soft, out-of-focus office in our palette."One image at the playhead, paid from your Prism AI balance

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

1
Paste the script and agree the scenes
One scene per sentence is the default; merge or split them before anything is built.
2
Build scene by scene
"Build scene 1" lands its layers. Check it, ask for the next.
3
Record and transcribe the voiceover
Record it in your own tool, put it on a layer, and ask for a transcript.
4
Time the scenes to the words
"Start each scene on its sentence" sets every in point; name a word to hold a scene longer.
5
Revise in words
A flat scene or a wrong colour is one sentence, and the same layers are rewritten.

Brand and style

Brand elementHow to ask
Coloursgive the hexes once: "background #F5F1E8, ink #1A1A1A, accent #FF3D05"
Typethe PostScript name, no spaces: Inter-Bold
Logo"import ~/brand/logo.svg as shape layers"
Motionname the feel: "soft overshoot", "iOS default ease", or a house ease
A change of mind"swap the accent everywhere" reaches every scene

What it needs

After Effects2024, 2025 or 2026, macOS or Windows
AI clientany MCP client: ChatGPT, Claude, Claude Code, Cursor, Codex and others
Plana Free account covers the build; transcription is a Prism AI capability paid from your balance

What it does not do

Prism does not generate a voiceover: record it, and the scenes are timed to it. There is no stock library: icons are shape layers or your own SVG, and a background plate can be generated one image per run, paid from your balance. It does not write the script.

Common questions