With AI native video tools, the bar for motion graphics just got a lot higher. We made this using our free Tesseract plugin for ChatGPT, based on a single reference video.

html could never

not a chance

I've already got it cooking with Codex right now. So far so good.

Wow! Going to try this out with other launch videos

this is insane man wow cannot believe my own eyes

damn

👀👀

BROOO 💀

🥵

Nice copy

this is insane everybody at mirage needs a raise, gonna be working with this a lot

html is scared

trembling

Guy who uses math to make video doubts computers could use math better than him. 🤣

So we can give it reference videos and it can replicate with this?

do you mean the workflow was "reference-to-motion"?

Holy moly!

tesseract plus gpt-6 astra combo is turning heads for good reason

Surprised @HeyGen, @LumaLabsAI, @Hailuo_AI, @FlowbyGoogle or @runwayml didn't build this first. Just goes to show how blind the "vision" model providers are.

Can you share a WIP of it?

What?

cool

Why only chatgpt?

Curious how much of the output stays editable after the gen. Production usually dies at cleanup, not at the first pass.

@DowloaderBot

Show us the prompt Cause here's what Tesseract made with my Astra subscription on max effort with the original video as reference And it sucks