Vox-style paper-collage explainers are everywhere right now.
They're pulling millions of views on TikTok and Instagram, and media buyers have started running them as paid ads. They're working. The style reads as editorial, not as an ad, so people actually watch.
Until recently, making one meant hiring a motion designer or spending days in After Effects.
Now Google Omni Flash generates this style shockingly well. Torn paper textures, cutout typography, hand-cut illustrations, narration baked into the video. And it generates fast.
I built a skill that automates the entire pipeline. You answer a few questions, and it produces the whole ad for you inside MaxFusion AI. No prompt copy-pasting, no babysitting.
Here's how the whole thing works.
What the skill does
You load the skill and tell it about your product.
It asks a short set of questions. Do you have a product image? Do you want a mascot on screen, or a faceless explainer? Vertical or horizontal? That's roughly it.
If you upload a product image, it gets attached as a real reference to every clip, so your product looks identical across the whole ad.
Example with Hevy app
Hevy app example
If you pick mascot mode, you upload a character image, and the skill converts it into a paper-cutout version that presents the ad to camera. If you pick faceless, you get the classic Vox explainer: no talking heads, just illustrations, labels, and arrows making the message land.
Then it writes the voice-over script and the full visual direction, shows you everything for one approval, and generates every clip in MaxFusion AI.
Each 10-second clip chains into the next using the last frame, so the ad flows as one continuous piece. You don't need to do anything as Claude does it all for you.
You come back to a finished ad.
And if you don't use MaxFusion AI: I'm also releasing a second version of the skill that skips the automation and hands you the finished prompts to copy-paste into your own tool. Same creative system, manual delivery.
Why Omni Flash and not Seedance
Other tools showing off this style run Seedance as their default. That's the wrong call. Especially for your pocket and time.
Seedance is more expensive per generation. It's also slower, and it's notoriously bad at 2 things this style depends on: words rendered on screen, and words spoken by the narrator.
A Vox explainer is built out of on-screen labels and narration. When the model garbles both, you regenerate. And regenerating means burning credits and tokens on clips you throw away.
Omni Flash holds text, holds speech, holds style consistency across clips, and finishes in a fraction of the time. For this specific format it's not close.
And if you need to regenerate it is cheaper but most importnatly so much faster.
Fixing the narrator voice
One quirk you'll hit: Omni generates the narration inside the video, and it doesn't guarantee the same voice across clips.
The skill already locks the gender, so you won't get a female narrator in clip 1 and a male in clip 2. But you can still end up with 3 clips that all have male voices, and each one sounds like a different guy.
Here's the fix:
- Stitch all the clips together in CapCut.
- Export the full video.
- Upload it to MaxFusion AI.
- Open the contextual menu and choose Replace Voice.
- Pick a voice from the library, or from your own cloned voices.
One consistent narrator across the entire ad, in one pass.
Where to run these
These aren't just organic content. The format converts as paid creative.
App marketers are using them to explain what the app does in 30 seconds. Supplement brands are using them to walk through the mechanism without a talking head. Any product that needs a "here's how this works" moment fits this format.
The editorial look buys you watch time that a standard UGC ad doesn't get.
So what different is this skill from others I have seen?
Other skills you seen posted, give you the prompts and tell you to go generate, go extract last frame and use it as reference, then generate again, go edit it together outside.
My skill is for Maxfusion AI MCP, and through claude you simply chat with it for 30 seconds and it does everything for you.
Go try it out - https://maxfusion.ai/?via=UGC-IG
I drop these skills and workflows in my free Telegram channel before they go live on X.
Join here: t.me/primegenai
Full YouTube breakdown soon.
