How I Made Professional 3D Motion Graphics With GPT 6 Astra (Full COURSE)

3D motion-graphics clips are everywhere on Instagram, TikTok and YouTube, and companies pay hundreds of dollars for a single one. Normally each takes hours in After Effects, one of the hardest programs to learn. I built a workflow where GPT-6 Astra controls After Effects and builds the animation: idea, script, visuals and motion from one master prompt, in under 15 minutes.

Step 1: Connect Higgsfield to ChatGPT

On Higgsfield (50% off) → MCP, select ChatGPT, click Add Higgsfield plugin, then Install plugin in ChatGPT.

Step 2: Give it After Effects

GPT can now talk to Higgsfield, but not to After Effects yet. On higgsfield.ai, scroll to Plugins → Adobe After Effects and download the installer (free with your ChatGPT subscription). Install it with After Effects already on your machine (Adobe offers a 7-day trial). Then in AE: Window → Extensions → Higgsfield. The connection is live.

Step 3: Run the master prompt in Codex

Switch from ChatGPT to Codex, paste the master prompt and send. You get five video ideas (or name your own). I picked The real reason you keep scrolling. GPT writes the script in five 6-second scenes (a 30-second clip) with the on-screen typography for each.

Step 4: Lock the style

Type next and choose a visual style: signature editorial (warm off-white paper), dark technical, or high contrast. GPT then builds one reusable style master composition (background, text styles and colors) so all scenes look designed by the same person. Copy the command it gives you, paste it back, and watch an empty AE project build itself layer by layer. Mine had exactly the 15 layers GPT promised.

Step 5: Hero assets

Type next. GPT writes an image prompt per scene, generates the visuals with Higgsfield, removes the backgrounds so they’re clean isolated objects, and imports them into AE automatically. No downloading, no dragging.

Step 6: Animate scene by scene

Type next, paste back the build command, and scene 1 appears in the timeline and animates itself, with sound effects. Codex shows a preview. Every layer, animation and keyframe stays fully editable. Repeat for scenes 2–5.

Step 7: Voice and quick edit

Generate the voiceover from GPT’s script with any TTS tool, like ElevenLabs. In CapCut or AE, add it to the clips, delete long pauses, sync to the animation, and add background music.

📝 Your notes on this video

Saved on this device only

Copy notesDownload .txt