We built a RENDER ENGINE for AI! Until now, AI-generated motion graphics have mostly been frontend code screen-recorded into video. A clever repurposing, but extremely limiting. Quality graphics requires optical effects, particles, 3D lighting and physics simulations, and cinematic post-processing. So we built a native GPU renderer from scratch,... designed for AI to write code directly at the rendering layer. No frontend frameworks. No Blender. No video generation models. Just code → pixels. Welcome to the real agentic era of motion graphics.show more

still need raytracing, motion blur, depth of field and less flickering on shadows. Good start!

How is it different from Tesseract from @trymirage

@grok so this is basically web 3d rendering?

Did you reverse engineer windows movie maker? 🥀✌🏼

This is a major leap for AI creativity. Direct GPU rendering opens up a whole new level of possibilities for motion graphics.

Where's the link to the repo?

Brilliant, ChatCut! Built render engine for AI, beyond screen-recorded frontend code, quality graphics sounds revolutionary and truly inspiring

Code straight to stunning pixels this is the future of motion graphics!

finally someone said the screen-recorded frontend part out loud. what's the hard bit for the agent now, getting the physics right or knowing when to stop adding particles?

😍😍😍😍😍😍😍😍

Code becomes cinematic pixels via GPU

No more Threejs? 😢

Brilliant leap from screen recordings to true native GPU rendering

This opens up incredible possibilities for AI motion design

Code to pixels directly is a huge leap for AI motion graphics

This is a huge leap for agentic motion graphics capabilities

This is a huge breakthrough for AI-powered motion graphics!

Voilà, this is the next generation.