@OpenAI Codex pretty much one-shotted this video with @HeyGen Plugin + HyperFrames (prompt in thread) CapCut probably never thought that it'd be threatened by vibe-coding. We got 4 significant launches around HyperFrames this week 🚢 the team is cooking hard rn!

Prompt: 15 seconds. LAURA ("Your Digital Twin") — grey blazer, dark framed glasses, dark brown hair — moves through the ancient Library of Alexandria. Towering shelves of papyrus scrolls stretching floor to ceiling. Ancient scholars working at tables in soft focus around her — reading by clay oil lamp light and window light. Camera tracks alongside her continuously. Medium close-up. No music. LAURA moves through the library as she teaches — already walking from the first frame. She trails one hand along a shelf of scrolls as she passes. She pulls a scroll partway off a shelf — glances at it — slides it back in and keeps walking. She passes a scholar hunched over a table reading by the light of a small clay oil lamp and briefly gestures toward his work without breaking her address to camera. Natural fluid movement throughout. She looks directly at camera the entire time. LAURA teaches directly to camera while moving. Exact words only: "The Library of Alexandria held an estimated 700,000 scrolls. It wasn't destroyed in a single fire — it declined slowly over centuries as funding dried up and scholars stopped coming. The real tragedy wasn't the burning. It was the forgetting." At 11 seconds — a wooden door already set into the library wall becomes visible ahead as she walks. It has always been there. She reaches it still looking at camera. Opens it — her warm office fully visible and established beyond. Bookshelves. Microphone. Window with garden. She walks through. Crosses to desk. Sits naturally into exact seated position from reference image. Looks at camera. Ready. Camera tracks continuously. Never cuts. Clear and stable facial features. Widescreen. Shallow depth of field. Film grain. 4K HD. No blur. No ghosting. No music.

Codex is now a video editor All you need is @HyperFrames_ + @HeyGen plugins Generate talking avatars, tts overlays and add animations with HyperFrames/stitch it all together The UX of the codex app is great for video editing, we've been playing with it internally a lot

That’s a wrap on our first @HeyGenDev Hackathon 33 teams built some insanely creative projects To give every submission the review it deserves, Demo Day is moving to Monday, May 18 • 1:30 PM PT Top demos + winners announced live Join us at this link:

The HeyGen Hackathon is officially underway 130 developers, both IRL and online, have 24 hrs to build scalable integrations on top of HeyGen’s video platform Prizes include: • $3K Grand Prize • $1K Best AI Agent Setup • $1K Best Use of HeyGen Who will take the crown? 👀

HeyGen is now built into Codex One click install Your A-Roll, B-Roll, audio, captions, and motion graphics all live in one workflow Call on your avatars, change looks, edit scenes, and generate videos without ever leaving Codex Big thanks to @OpenAIDevs for the support

We’ve tested a lot of Avatar Shots workflows The biggest difference between average AI videos and great ones usually comes down to direction • Scene structure • Camera movement • Environment prompting • Character behavior Here’s the workflow we keep coming back to ↓

I wanted to see how far I could push character consistency in AI Video Generator in @HeyGen So I reused a character sheet from another project and added reference images to create an animated cat food commercial with consistent packaging, styling, and shots Breakdown + prompts

You can now render @HyperFrames_ on @Vercel One click gets you a full programmatic video pipeline: live preview, server-side rendering, and MP4 output All on Vercel infra Deploy instantly Build your hyperframes app with vercel 👇