Big Idea

The argument in one line.

Three tools -- MaxFusion for research and orchestration, Claude for script adaptation, and Seedance 2.0 for generation -- let you remake any winning UGC ad format for any product in roughly five minutes per video.

Who This Is For

Read if. Skip if.

READ IF YOU ARE…

You run paid social ads and want to test more UGC creative without hiring actors or a production crew.
You manage products on Amazon, Shopify, or DTC brands and need a steady pipeline of short-form video ads.
You are already using Claude for content work and want to extend it into video ad production.
You want to offer AI UGC production as a service to clients and need a repeatable, scalable workflow.

SKIP IF…

You need authentic human UGC -- AI-generated faces still read as synthetic to audiences who look closely.
You are not running paid ads and have no product to promote -- this workflow is ad-production-first, not organic content creation.

TL;DR

The full version, fast.

MaxFusion is a node-based workflow builder for AI ad production. Its Research node surfaces top-performing TikTok and Meta ad videos by views and engagement; its Content Analyzer deconstructs the structure; its Video Generator powered by Seedance 2.0 remakes the video in minutes. Claude fills the gap between formats: paste the content analyzer output, describe your product, and Claude rewrites the script to fit your offer. For more complex multi-person formats, Claude Cowork with a custom UGC Decoder skill reads a reference video and outputs scene-by-scene prompts; GPT Image 2 generates the actor stills; Seedance 2.0 animates each one; MaxFusion Video Editor stitches the final sequence. Total time per video: five minutes for simple clones, under an hour for multi-scene builds.

Free for members

Chat with this breakdown — free.

Sign in and you get 23 free chat messages on us — ask for the hook, quote a framework, find the exact transcript moment, generate a markdown action plan. Bring your own key when you want unlimited.

Create a free account →

Chapters

Where the time goes.

00:00 – 00:27

01 · Introduction

Side-by-side comparison hook; promise of free prompts, skills, and workflows.

00:27 – 01:28

02 · Find Top Performing UGC Videos

MaxFusion Research node -- TikTok and Meta Ads Library search; filter by likes/views; select a winning video.

01:28 – 02:59

03 · Clone Any Video With One Prompt

Content Analyzer deconstructs the selected video; Video Generator with Seedance 2.0 remakes it; result shown side-by-side.

02:59 – 04:26

04 · Change the Product in the Video

Claude prompt takes content analyzer output plus new product description; adapted script returned to MaxFusion with product image reference; new video generated.

04:26 – 08:54

05 · Build Multi-Scene UGC From Scratch

Claude Cowork + UGC Decoder skill analyzes a reference format; per-actor prompts generated; GPT Image 2 creates actor stills; Seedance 2.0 animates each scene; MaxFusion Video Editor assembles final cut.

Atomic Insights

Lines worth screenshotting.

MaxFusion Content Analyzer deconstructs any winning UGC video into a structured brief you can hand directly to Claude or Seedance.
Seedance 2.0 reproduces visual style, energy, and pacing from a reference -- not just the script, the whole aesthetic.
Claude role in this pipeline is narrow: take the content analyzer output, swap in the new product, keep the proven structure.
The Meta Ads Library longest-runtime filter is a profitability signal -- if a brand is still running it, it is converting.
A Claude Cowork skill can analyze a reference video and output per-scene animation prompts, replacing hours of manual shot-listing.
GPT Image 2 generates photorealistic actor stills at 9:16 that Seedance 2.0 can animate into ten-second video scenes.
MaxFusion supports 35+ languages, so a winning English-market ad can be localized and rebuilt for any regional market in the same workflow.
The five-minute-per-video claim is credible for one-clip clones; multi-scene builds with six or more actors take around thirty minutes of generation plus assembly.
The car dealership prediction format -- each salesperson guesses how many cars they will sell today -- is a high-engagement template because per-person predictions create natural variation and rewatch hooks.
This workflow commoditizes UGC ad production: the competitive advantage shifts from access to actors to quality of research and offer clarity.

Takeaway

Three tools that collapse UGC ad production to minutes.

WHAT TO LEARN

The research-analyze-generate pipeline eliminates the two biggest bottlenecks in UGC ad production: finding what already works and casting humans to perform it.

Filter the Meta Ads Library by longest runtime to identify ads still converting -- brands do not keep paying for creative that is not working.
A content analyzer that deconstructs a video structure lets you adapt a proven format to any product without reverse-engineering it manually.
Claude role in this pipeline is narrow: take a structured brief from the content analyzer, substitute the new product, and preserve the original format proven hooks.
Seedance 2.0 reproduces the aesthetic feel of a reference video -- not just the script -- which matters because visual tone drives scroll-stop behavior.
Multi-scene AI actor formats require a scene-planning step before generation; Claude Cowork handles this via a reference video upload.
GPT Image 2 at 9:16 generates photorealistic actor stills precise enough for Seedance 2.0 to animate convincingly -- the two tools together function as a casting replacement for simple UGC scripts.
Language localization is a direct reach multiplier: one winning English-market ad rebuilt in 35+ languages via MaxFusion expands reach without additional creative development.

Glossary

Terms worth knowing.

UGC ad: A short-form video ad styled to look like organic user content -- handheld camera, natural speech, personal recommendation -- rather than a polished brand commercial. Performs well on TikTok and Meta because it blends into feeds and earns trust.
MaxFusion: A node-based AI ad production platform connecting research (TikTok and Meta Ads Library), content analysis, image generation, video generation, and editing into a single visual workflow.
Seedance 2.0: A video generation model available inside MaxFusion that animates still images or generates video from text prompts, capable of reproducing a reference video visual style and energy.
Content Analyzer node: A MaxFusion workflow node that takes a selected video and outputs a structured breakdown of its format, script, visual style, and pacing -- used as input to the video generator or Claude.
Claude Cowork: A Claude interface mode that accepts file uploads (images, video) directly into the conversation, enabling Claude to analyze visual content and generate structured prompts from it.
UGC Decoder skill: A custom Claude skill that ingests a reference video and outputs a scene-by-scene production plan including actor descriptions, dialogue, and animation prompts.

Resources

Things they pointed at.

01:00toolMaxFusion ↗

04:26linkClaude Skills + workflows download ↗

06:30toolGPT Image 2

Quotables

Lines you could clip.

01:58

“It literally recreated the exact same style, energy, and overall vibe of the original video.”

Proof moment -- before/after result stated cleanly→ IG reel cold open↗ Tweet quote

04:07

“We took the casual car review format from a winning ad and applied it to a completely different product, just couple of clicks.”

Concise summary of the core value prop→ TikTok hook↗ Tweet quote

08:15

“Think about how long it would take to produce something like this. Now you can do the whole thing in under one hour.”

Time-compression contrast -- strong shareability→ Newsletter pull-quote↗ Tweet quote

The Script

Word for word.

Read-along

Don't just watch it. Burn it in.

See every word as it's spoken — crank it to 2× and still catch all of it. The same dual-channel trick behind Amazon's Kindle + Audible.

UGC videos like these are blowing up right now, and brands are using them to get thousands of new customers every single day. So I built a system using Claude and Sea Dance two point o that lets you find winning UGC videos and remake them in minutes. Today, I'll show exactly how it works by creating multiple different AI UGC videos from scratch and giving away all the prompts, Claude skills, and workflows completely free.

And all I ask in return is for you to drop a like on this video. When I make UGC videos like this one, for example, the first thing I always do is find what is already performing well. And this used to take me hours.

But Max Fusion, the main software I used to create AI UGC videos, now lets you find the best performing UGC content in any niche, and then you can literally select any video and remake it with one click. Let me show you how it works because I know it sounds too good to be true. I leave a link for Max Fusion in the description below if you want to follow along.

But once you log in, click new project, select flows at the top, and create a new flow. Click the plus icon and select research. From here, you can choose the meta ads library or TikTok content to find best performing UGC videos.

Let me go with TikTok and type sweater review.

You can filter by country, sorting settings, and set likes and views thresholds. I will leave it as it is and hit search. And just like that, we have a library of the best performing videos with views, likes, comments, and shares all visible right here.

Now here's the crazy part. Select the video you want, click save, then tap the plus icon and add the content analyzer note. Connect them together and run it.

In less than a minute, it breaks down the entire video structure. Then you add a video generator, connect the content analyzer node, make sure students two point o is selected, and hit generate. And here's the result.

Just got this crazy pretty sweater, and this one is fully embroidered from front to back. I just got this crazy pretty sweater, and it is fully embroidered from front to back. It literally recreated the exact same style, energy, and overall vibe of the original video.

And in a minute, I'll show you how to swap this for any product you want and make longer videos as well. Now here we can also find the best performing ads that are being run on Meta. Just add the Meta ads research node.

You can select the country the ads are running in and filter the longest running ones. That way you know that the ad is profitable. So, for example, I would type in electrolytes and I can find all of the different ads that are running for electrolyte products.

You can just scroll through and find the video that you like, select it, run it through the content analyzer and video generator. And in about five minutes, I get this exact video remake. And to drink, we have some watermelon candy electrolytes with no mustard.

And to drink, we have some watermelon candy electrolytes with no mustard. If you are new and you're looking for the electrolytes that I drink, this is what they look like. You If are new and you are looking for the electrolytes that I drink, this is what they look like.

Yeah. So it even picked up the exact talking points and the whole aesthetic from the original ad. Now let's say you like this review style video for this electrolyte drink, but you're selling a different product.

This is where Claude is going to help us. So go to Claude and use this prompt. Just paste in the content analyzer output and describe your product in the brackets right here.

And it takes the original video structure and rebuilds it for your product. Then we take this prompt back into Max Fusion Flows. Click the plus icon and select prompt.

Click plus again and select video generator. Click the plus icon one more time and select upload media. Drop in an image of your product.

Paste the cloth prompt into the prompt note. Connect the prompt note and the product reference image into the video generator. Make sure Sea Dance two point o is selected with nine by 16 format, and just let it run.

And, again, in less than five minutes, we get this. This is the water bottle I bring to the gym every single day, and I am not switching. The texture on this thing is so satisfying.

I actually look forward to taking it out of my bag. It fits in every gym bag I own and it looks so clean sitting on the floor next to my mat. Honestly, just the nicest bottle I have used.

You can find it on Amazon. Just search gym water bottle, and it will come up. So as you can see, we took the casual car review format from a winning ad and applied it to a completely different product, just couple of clicks.

And each of these took me about five minutes to make. So you can see how using this workflow, you can really crank out a ton of UGC videos every day. This is the power of combining Max Fusion research with Claude, and this also works in most languages.

You can take winning ad from any market and rebuild it in English, Spanish, French. Max Fusion supports 35 plus different languages. Now let's say you want to rebuild this video right here, which is much longer.

It also involves different people and has more different scenes. This is where I built these cloth skills to help you out. So first, install the cloth skills that we're going to use, and you can download them from the description below.

Extract them, then go to cloth, click customize, then skills, click the plus icon, create a skill, and upload them one by one.

And that's it. Now these skills are built in a way that allow you to drop a reference video into cloth. They analyze it and build a whole system to remake that video from scratch using CDN two point o.

So I found this video right here using the research tool on Max Fusion. The whole format is about asking each employee how many of a product they think they will sell today. Each person gives their own prediction on camera.

This is a brilliant format for any business, but creating content like this takes a lot of time and setup. So instead, I'll show you how to build this entire thing using AI. And it works for any product too.

Supplements, skincare, clothing apps. You can just swap the employees for brand ambassadors, and the concept stays the same.

I will use this format for a car dealership where each salesperson predicts how many cars they will sell today. So after installing the skills, I open Cloth Cowork and activate the UGC decoder skill.

The reason I'm using Cowork is because it lets me upload the reference video directly into Cloth. I drop in the reference video together with this prompt and click send. The only thing you need to adjust in this prompt is your business type and product.

Now Claude reads through the entire video and within minutes, it breaks down the whole structure. Now this might look overwhelming at first, but don't worry. We just follow the steps one by one.

The first thing we need is our actors, the salespeople who will be making the predictions in the video. So I asked Claude to give me the prompts to generate each chapter. I use this prompt right here.

And by the way, all of these prompts that you see me using are in the description down below as well together with cloth skills and all of the workflows. Now let's go to Max Fusion and open a new flow. Click the plus icon and add a prompt note, then click plus again and add an image generator note.

Connect them together, drop in the first actor prompt, select GPT image two, set the format to nine by 16, set the quality to high, and hit generate. And just like that, we have our first actor ready. I will repeat this for all of the remaining actors, and here are all of the salespeople ready to go.

Now we just need to animate each one of these and turn them into actual video scenes. For that, I go back to Claude Coburg and drop in the screenshots of all of the generated actors along with this prompt. Now let's go back to Max Fusion Flows and animate each actor.

Click the plus icon, select prompt, then plus again, and select video generator. Connect the prompt and the generated actor image to video generator. Drop in the first animation prompt.

Make sure c d r n's two point o is selected. Set it to ten seconds and nine by 16 format, and just hit generate. So here is our first scene.

Hi. I'm Ashley. I'm the receptionist here.

And, honestly, I think we're gonna sell at least six cars today. Now I will repeat this for all of the remaining actors and generate all of the remaining scenes. And in about thirty minutes, I have all of the remaining scenes right here.

Now I just click plus icon again, select video editor, connect all of the generated videos into this note, and we can quickly arrange the scenes in the right order. After assembling the scenes, I save the final video on my computer, add captions in my editing software, and we have the final video ready. Ray, lot manager.

Twenty two years on this lot, and today, we're moving 20 units. I don't guess. I know.

Marcus, I just started three months ago. I'm saying three cars today, but watch. I'll probably do four.

Think about how long it would take to produce something like this. Now you can do the whole thing in under one hour. These videos are great for organic content and also could be run as ads.

The exact flow that I built for this along with all of the cloth skills and the link for Max Fusion are in the description below. AI UGC videos are getting better and better each week. All thanks to the tools like Max Fusion and the team behind it that constantly makes the software better.

They also recently updated their credit system, so now you get significantly more video generations every month for the same price. And very soon, they're planning to have an AI agent that will be able to create these workflows for you automatically.

If you want to see how I make viral UGC carousels using cloth and GPT image too, you can watch that video by clicking here. Thank you so much for watching, and I'll see you in the next

The Hook

The bait, then the rug-pull.

Split-screen on the open: a real UGC sweater review on the left, an AI-generated remake on the right -- nearly identical energy, completely synthetic. The promise lands before a single tool is named.

Frameworks