Big Idea

The argument in one line.

AI motion graphics that look like you come from a personal system of templates, rules and a skill, not from a better prompt or a paid animation tool.

Who This Is For

Read if. Skip if.

READ IF YOU ARE…

A YouTuber who wants animated explainer boards synced to their narration without paying an editor $500 or more per video.
An editor or freelancer who wants to add motion graphics as a premium service on top of existing edits.
Someone using Claude Code who wants a concrete pattern for packaging taste and rules into a reusable skill.

SKIP IF…

You want a one-click tool and have no interest in building or maintaining your own templates.
You only need short-form captions, since the system shown targets long-form explainer boards.

TL;DR

The full version, fast.

Animated boards are just HTML pages driven by GSAP, and what makes them good is the skill behind them, not the prompt. Sirio builds a style from a references folder in Claude Design, fixes the canvas in exact pixels, records every correction as a rule in a checks file, and builds a template library of nearly 300 boards. He packages it all into a skill, then uses Claude Code to transcribe with Whisper, plan boards, sync them to each spoken word, and render via Chrome and FFmpeg. He closes with ways to earn from it: saving editor fees, selling a done-for-you system, templates or an MCP.

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 – 02:00

01 · Hook and promise

Sirio promises a free pack of 300 Claude-built boards and previews the nine-step system.

02:00 – 04:28

02 · Three myths

Breaks the Katana, better-prompt and need-a-designer myths.

04:28 – 05:52

03 · Collect references and find your style

Steps 1 and 2: reference folder, then locking a look in Claude Design.

05:52 – 07:04

04 · Set the canvas

Step 3: exact pixel boxes, padding, safe zones and end screen.

07:04 – 08:46

05 · Fine-tune with checks and QA

Step 4: a checks file of do and do-not rules plus an automated QA prompt.

08:46 – 11:26

06 · Build the template library

Step 5: nearly 300 boards, 200 plus templates, all HTML and GSAP.

11:26 – 13:25

07 · Turn it into a skill

Step 6: files for style, layouts, checks, motion and a board catalog.

13:25 – 15:35

08 · Turn audio into boards

Step 7: Whisper transcript and the board brief skill in Claude Code.

15:35 – 17:45

09 · Sync to voice and render

Steps 8 and 9: word-level sync, Chrome frame capture, FFmpeg, MKV parts.

17:45 – 21:45

10 · How to make money with it

Saving editor fees, prospecting creators, selling templates, SaaS or MCP.

21:45 – 23:19

11 · Close and community

Productize your expertise, start with one board, join Public AI.

Atomic Insights

Lines worth screenshotting.

A feature like Katana is a skill with many instructions and examples locked behind an API, and the animation itself is just an HTML page.
Giving Claude a skill with templates produced an animated drawing of the idea where a plain premium-motion-graphic prompt produced a slide with a title.
Tell the AI to keep a 480 by 360 box empty in the bottom-right corner. Saying leave room for my face gives it nothing to measure.
Every correction to a board should become a do or do not rule in one checks file, so the same mistake never happens twice.
AI does good and fast work when it has many categorized template examples, so build one template per effect you like.
Write board catalog tags in the words you actually say aloud, so matching a spoken line to the right board is easy.
Telling Claude to export HTML to MP4 loses the CSS. Capturing each frame in Chrome and assembling with FFmpeg keeps the design exact.
Save render parts as MKV, not MP4, because an interrupted MKV still opens while a cut-off MP4 is lost.
A 20-minute video at 30 fps is about 36,000 images, so split it across parallel workers and test 60 seconds first.
Long-form motion boards can bring about $300 per video, while short-form work sells for roughly $50 to $100.
One long-form client posting weekly is about $1,200 a month, and three such clients reach about $3,600.
You are not selling designs to a creator, you are selling a system with their style, templates and skill.

Takeaway

Your style lives in a skill, not a prompt

WHAT TO LEARN

Good AI motion graphics come from templates, exact numbers and accumulated rules packaged into a skill, and that system can also be sold.

01Hook and promise

Motion graphics that look generic usually come from everyone using the same tool, so a style that is recognizably yours is the differentiator.
A board-based system built over two weeks can replace an editor charging $500 or more per video.

02Three myths

An animation tool like Higgsfield's Katana is just an HTML page driven by a skill, and the skill's instructions are what you are actually paying for.
What the AI already knows (templates, examples) beats a better prompt: the same line produced a plain slide without a skill and an animated drawing with one.
You need taste, not After Effects skills. Knowing what is good matters more than knowing keyframes.

03Collect references and find your style

Start by saving screen recordings and Pinterest UIs into one references folder so the AI has real taste to draw from.
Tell the design tool your use case and feel, ask it to combine your references into something unique, and iterate for days or weeks until it feels like you.

04Give the AI the canvas in numbers

Vague instructions like leave room for my face fail. Specify exact sizes such as a 480 by 360 box in the bottom-right corner.
Define padding, safe zones and the end screen once in a template so every new video starts with the layout locked.

05Checks file and QA loop

Keep one checks file where every correction becomes a do or do not rule, so the same mistake never repeats.
Having the AI run a QA prompt over every board catches most mistakes before you do and turns them into more rules.

06Build a template library

AI works fast and well when it has many categorized examples, so build a template once for every effect, chart and text animation you like.
The stack is plain HTML plus GSAP, with SplitText for word-level text motion, which keeps every template editable and free.
Tag each board with a description of what it does, such as correcting a belief, so the agent can match it to what you say.

07Turn it into a skill

A skill is a markdown file set: style, layouts and face box, checks, motion timings, and a one-line-per-board catalog.
Write catalog tags in the words you actually say in videos, so matching a spoken line to the right board works reliably.

08Turn audio into boards

Transcribe with Whisper at word level so the AI knows the exact second each word is spoken.
A board brief asks of every line whether it can be understood by listening alone, and numbers, prices, comparisons and steps get a board, one idea each.

09Sync and render

Sync each animation to start on the word that names it, and cap how long anything sits still.
Export by opening each board in Chrome, capturing every frame and assembling with FFmpeg. Screen recording drops frames and HTML-to-MP4 shortcuts lose the CSS.
Render in parallel workers, save parts as MKV so a crash does not corrupt them, test 60 seconds first, and verify the frame count at the end.

10How to make money with it

Saving an editor fee is the first return, and editors can add motion graphics on top of existing edits.
Pitch 20 creators without motion graphics by making three free boards from their last video, and sell a system built around their channel rather than designs.
Long form supports about $300 per video while short form is about $50 to $100, and templates or a skill-as-MCP can be sold on top.

11Close

Start with one board, turn every failure into a rule, and keep going for a week. The value is the thinking packaged into the system, not the tool.

Glossary

Terms worth knowing.

Board: A single animated HTML visual that explains one idea from the script while the creator speaks.
GSAP: A free JavaScript animation library used to build timelines for the HTML boards, with a SplitText plugin for animating words and lines.
Checks file: A running document of do and do-not rules, each added after a correction, which the AI follows on every later board.
Skill: A set of markdown instruction files that give Claude a repeatable workflow, here covering style, layouts, rules, motion and a board catalog.
Board brief: A plan generated from the transcript that groups lines into one-idea boards and lists what goes on each.
MKV: A video container that stays playable if a render is interrupted, unlike a half-written MP4.
MCP: A protocol that lets a user plug a skill and its assets into their own Claude, which can be sold as a subscription.

Resources

Things they pointed at.

02:00productKatana (Higgsfield)

04:30toolClaude Design

10:00toolGSAP

14:00toolWhisper

16:20toolFFmpeg

20:50productGumroad / Stan Store

22:38linkPublic AI community ↗

Quotables

Lines you could clip.

03:34

“The prompt I believe matters less than you think. What matters is what AI already knows before you ask.”

contrarian one-liner→ TikTok hook↗ Tweet quote

21:46

“Productize your expertise.”

short repeatable line→ 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.

Today you will learn how to become an expert in AI motion design so that you can make a sh** ton of money. And no, we are not going to use Katana from Hicksville. We're actually building our own and I'm going to walk you through it step by step because right now everyone's AI motion graphics are looking exactly the same and I want yours to look like you.

I made this boards with Claude and it didn't cost me anything except my Claude subscription. I'm on the $200 plan and I mostly use Opus 5 .5 on Medium. Now an editor would charge you at least $500 to do something similar.

I built 300 of these templates in two weeks, all in my own style. And today I'm going to give it all away for free. The boards, the prompts, everything.

And at the end of this video, I'm going to ask you to stop watching videos like this one, mine included. Hey friend, welcome back. If you're new here, my name is Sirio.

I architect AI workflows and systems. I deploy APIs and productized features. And I build AI businesses, particularly in the media space.

And now I'm using these AI motion graphics on every single YouTube video, including this one. Now there are three kinds of motion graphics. Motion graphics for talking head videos, for faceless videos, and also for long form explainer videos, like this one, where the boards explain what I'm saying while I'm talking.

And that's what we're going to cover today. Meanwhile, take a look at this no face video animation. I've only used code for it.

A Boeing 737 can take off weighing over 80 tons, and the only thing holding it up is two wings and the air. So what are those wings actually pushing on? They aren't floating on the air, they're bending it down.

You were probably told the air on top has to hurry to meet the air underneath. But that isn't true. When engineers at Cambridge filmed puffs of smoke over a wing, the air on top got to the back first.

So everything we're going to talk about today, the boards, the prompts, the animations, everything you're going to find it for free on my GitHub. The link is in the description. And here's what you're going to learn.

First, I'm going to quickly break three myths. Then I'm going to show you how to get inspired, find your own style and build animated boards for your videos. how to fine -tune them until they feel like you, and how to turn them into templates and put all of it in your own skill so that you can use the whole system yourself or sell it if you wish.

So first, let's quickly break all the three myths. Myth number one is that you need a tool like Katana from Hicksfield to animate things. And I get it.

It's easy. There's nothing for you to set up because someone already has built the system for you, right? That's what you're paying for, actually.

The thinking that they did. Now, I haven't seen their code, but a feature like Katana is a skill. It has tons of instructions and examples so that AI knows what to make and how to make it look good.

Then they lock that skill behind an API and turn it into an NPC so that you have to pay to use it, right? Simple words. What's happening is that the animation that you're seeing is just a web page.

It's an HTML. And the part that makes it good is actually the instructions behind it. So the skill.

Now, if you want something that's fast and generic, sure, pay for the tool. That's fine. But if you want it to look like you and also have some leverage in this industry, guys, build your own.

Because everything that you are sold today. You can just build it yourself. Myth number two is that a better prompt gets you better graphics.

I tested this. I gave Claude the same line for my script two times with the same model. We used Opus 5 .5.

The first time, I simply said, make this a premium motion graphic. The second time, I gave it a skill. And now my skill has a bunch of templates inside, templates that I built myself, so Claude could look into them.

And this is exactly what it looks like. Look at the first one. It's basically a slide, a title, and some text.

And a nice background. Obviously, you can go further and extend your prompt and be more specific. Now, are you going to really get better results?

Look at the second one. It actually draws the idea and every piece shows up right when I say it. So the prompt.

I believe matters less than you think. What I believe that matters is what AI already knows before you ask all these things that you're trying to design. Myth number three is that you need to be a designer or know After Effects.

You do not have to be a designer. But you do need to have taste. You need to understand what's good.

You don't need to know how to use keyframes. You just need to have this visual eye. Okay, so now that we've talked about these three myths, what are we going to build today?

We're going into claw design, and I'm going to show you how I build my motion graphics for my use case, which is... YouTube videos, this video that you're watching right now. And just so you know, I'm not going to use Cloud Motion here.

I built all of this with Cloud and in Cloud Design, but Cloud Motion is available today. And I think that you can actually take this exact workflow and make it even better with Cloud Motion. I haven't tested it myself just yet.

And you can use this pipeline if you sell a course, short form, your presentations. There are so many use cases for this. And we're going to do this in nine steps.

The first six steps, we're going to build the system. And the last three, we're going to use the system that we build on a new video so how do we start well first of all we need to start by collecting ideas like any good artist like any good creator does we need to get inspired so that we can start building our boards so what you can do is you can screen record your animations that you like you can save the UIs that you like on Pinterest and put it all in a folder on your computer and we can call this folder something like references cool that's step one step two We have to find your style.

So how do your boards look and feel? We're going to open Claw Design, and you're going to drop in the reference folder that we just built. Then you tell it your use case, because every use case needs different kinds of boards, different dimensions.

And for me, I want boards that explain something while I'm talking. Then you'll tell it how you want it to feel. And I always tell it something like, take the ideas from my references, in this case my folder, and suggest how I can create something unique by combining everything together.

Then you'll lock in your look. Cloud design is going to give you a few options. You have to go back and forth until it actually feels like you.

Now, this could take you days or weeks, okay? Do not rush this process. I cannot tell you what's good for you, but I can tell you how you can get started.

Step number three is the canvas. And this is important. This is where everything goes on screen.

And in here, you need to give AI numbers. What do I mean by that? Well, because if you say, leave some room for my face as a placeholder, the AI does not know how much.

room you mean but you have to be very specific and you have to tell like keep this 480 by by 360 box empty in the bottom right corner so that you can get it right every single time for example this is what uh one of my templates looks like before i start any video it already knows exactly where the placeholder is and exactly where my face is going and i have some other templates where i define the padding so for example 60 pixels on each side the safe zones where nothing important ever goes there, and even my end screen with placeholders for the videos that I'm suggesting at the very end of my YouTube video with my subscribe button as well.

Step number four. After we've gone through this process and we have our style and canvas locked in, claw design. already gave us our first boards.

So this is what they look like. This was what mine looked like in the beginning. And what we're going to do is that we're going to start fine tuning them.

We are going to go through every board and look for what doesn't feel like us. And this is important. Before you start fixing anything, you need a place to save everything that you learn, like a little database for what I call your checks.

So the first thing I do is that I ask Claw Design to create an empty file for my checks. Every time I correct a board, add it there as a do or do not. And from there, every fix gets saved in that one file.

And for me, it was very specific. My first board looked very nice, but it didn't explain anything. There was too much text.

So every time I would tell it something like this, there's too much text on the board, draw it instead and make sure that it's showing one central idea, fix it and add a rule so this never happens again. And that's how you end up with a few rules. And these are some of the rules that...

I actually came up with while doing this fine tuning process. You can read and screenshot all of these, but they're all going to be in the GitHub for you. Now, a hack that I normally use to automate this entire process is to let AI start finding what's wrong and start testing it itself.

And I call it the QA, quality assurance. And this is the exact QA prompt that I use. You can see it on screen.

You can run this on every board and it starts to catch most of the mistakes for you. So every mistake turns into a rule that Claude records and I call them checks so that in the end you can package all of them into one skill. Step number five.

Okay, this is one of the most important things that I've learned while building this system, which is that if you want AI to do a good job and a fast job. Both. You need to give it a lot of examples.

And what are these examples exactly? Well, in this case, they're just templates. So you have to create templates for literally everything.

And the more, the merrier. And they have to be categorized right inside Claw Design, almost like a library of every single component or idea that you're going to use. And honestly, I'm making this my mission to create as many templates as I can for my use case.

So when I record a new video that maybe has a very different concept, I don't really have to reinvent. the wheel because pretty much everything is ready in the templates to be used. And look at this.

This is my current library and cloud design is almost 300 boards and over 200 of them are templates and every template gets a simple name so we can categorize them and use it in every video. So let me show you a few. These are my premium motion templates.

Every effect that I like, I built once as a template. So I never have to explain to AI again when we create the animations for my video. Then here are the charts.

Bar charts, line charts, waterfall. And every one of them already knows how to move in exactly what animations to use. And here I have some motion templates that define how each component animates.

I have text animations. keyframes, transitions, and a lot more. And if you're actually wondering how this works, as I said before, it's just all HTML plus a free animation library called GSAP.

And for text animation, GSAP has a plugin that's called Split Text, which splits the title into lines or words, so each can move on its own. Here's a bunch of keywords I use for my animations, so you can copy them straight into your prompt. Line chart.

Exploded layers. a video wall to hero, before and after slider, a scribble fill, a scribble highlight, a line draw, a virtual camera, stagger and collapse, keyframes, and a focus pull. If I say in the video, people think this, but it's actually that, that's a myth versus fact board.

This is the board right here. And if I say it repeats until it works, It's a loop.

So how does it know? Well, as you can see in the board, there's a section on the right where it literally describes the animation. And it has a tag, like correcting a belief.

So if something I say in my video matches the tag, the agent finds the right board because it goes through all of them. It goes through all the script. And that's something very smart you should start implementing into your workflow.

It's all this syncing between your boards, your designs, and what you were saying. Step number six. Let's make our own skill.

Now we take everything we just did and put that into a skill. And a skill, as you know, is basically just an MD file, which is a simple text file. And to do all of this, you only need to use Claude whenever it needs a tool like Whisper for your audio or GSAP for the animations.

It finds it, it sets it up for you, and does the work. The only thing you actually build yourself is the skill and the steps that we've covered so far. So your rules, your designs, your examples, and your templates.

And the reason we did everything until step number six is so that we can fit it to the AI to build a skill for us. And this is what my skill actually... looks like it's one main file and then a few other smaller files one for my colors and fonts one for my layouts and my face box one with all the do's and do nots so all my checks one for motion with every move and how long things take and one called the board catalog which is every single board one line each what it's for and when to use it and claude writes most of it for you straight from the board so when you're ready to turn everything into a skill This is what I do.

I always tell it something like this. Turn everything we did into a skill. First, write down my style, my canvas, and every rule in my checks file.

Then open every HTML board and log it in a table. The name, what it shows, what kind of sentence it's for, how it moves, and its tags. Write the tags in the words that actually say in the video.

So when we sync it to my voice, you can find the right board. And if two boards do the same job, tell me which one to keep. Because later when we sync all of this to your voice, the AI needs to find the right board.

Essentially, what we're doing right now is that we're turning all these visuals into text so that AI can actually read it. And obviously, when we use it to sync to our voice, it's not going to use the exact same words every time. It's going to draw an inspiration from it and change it based on the context of what I'm saying in the video.

That's your katana, my friend. Literally, that's what we are building right now. Okay, cool.

So let's do a quick overview. We got inspired. We found our style, we set it up in our canvas, we fine -tuned it, we made templates, and now we put everything into a skill.

So we built this whole system and it's time to actually start using it to create brand new boards for a brand new video. But before we use that, you need to get everything out of Cloud Design. Remember, we're in Cloud Design right now.

And all you have to do is just tell it to download my boards as an HTML file, plus every MD file that you've ever saved into one folder so that I can use everything in Cloud Code. That's what you say. And it will give you something like this, which you can download.

Step number seven, turn your audio into boards. So now we go to Clod Code. As I said, up until now, everything was in Clod Design.

Clod Code is the coding side of Clod. So it can work with all the files on your computer. You open a new project in Cloud Code and put two things in it.

First, the folder that we just downloaded with all the boards, your MD files and your skills and your new video or just the audio. The first thing to do is ask Cloud Code to transcribe it so that he knows exactly what the video is about. And you say something like transcribe my video with Whisper with timestamps for every word.

Whisper is a free tool and it gives Cloud every word that you say in your video or audio and the exact second that you say it. Then once everything. transcribed you run a skill that's called the board brief it's just another skill that i build and it's in the github too think of it like a plan for your graphics before anything gets designed it reads a transcript that whisper gave you and for every line it asks one question Can someone understand this just by listening?

If it's a number, a price, a comparison, or a list of steps, the answer is no. Then it groups those lines into boards, one idea per board, and writes down exactly what goes on each. At the end, you get a very simple file that says these are the boards this video needs, and this is what's on each board.

Then it checks with your own skill, the one we made from claw design. It reads the tags in that table, and that's how it finds which of your templates fits each line. And it builds your new board space on what you actually said.

Step number eight. Sync it to your voice. Okay, so now that we have our new boards as HTMLs, they are ready to animate.

It's time to sync them to what you're saying. And you do this in the same Clod code project. Remember, your transcript already has the exact second that you say every word.

So Clod looks at when you say a word, and it makes the animation for that word start at that exact same second. So the prompt is simply like... Sync every board to my transcript, everything shows up on the word that names it, and nothing sits still for more than, I said 5 seconds, you can say however many seconds that you want.

So that you get something like this. It's pretty cool. Now, if everything looks good, we have to export this entire thing.

And the issue is that this is an HTML. So even though it might be playing for you right now, it's not an actual video. You cannot just put it in your like editing platform or you cannot upload it anywhere.

So we have to turn this into an MP4. How do we do that? Here's something I learned the hard way so you do not have to do what I did.

I thought that I could just tell Claude to export an HTML to like an MP4 and that would look exactly like it does in the browser. That's not the case. It didn't keep the same.

design style did not keep the CSS so you have to tell it exactly this open each board into Chrome record every single frame and then put all those frames together into a video with FFmpeg with my voice FFmpeg is a free tool and that's the only way it gives the exact same design that you see in the browser and there's no screen recording involved here whatsoever because if you're gonna do that then it's gonna drop the frames Everything runs locally on your own computer and it renders at 30 frames per second.

So a 20 minute video is about 36 ,000 pictures in this case. To make this faster, Claude splits the video into parts and renders them all at the same time. And these are called workers.

I use six workers in this case. Normally a 20 minute video takes about 30 minutes to render. Pro tip, always render 60 seconds first before you start a 20 minute video render.

Save each part as an MKV file, not an MP4 because if the render gets stuck or gets... It's cut halfway. An MP4 is just going to be broken and you lose all the render, like all the data.

But an MKV still can open. And at the very end, Claude checks that the number of frames match the length of the video. So the prompt I use for that is something like this that you can see on the screen.

And you guys listen to me for a second because I'm going to turn into a coach right now. The only thing that you need is quite literally Claude. Nothing else.

Claude and you actually thinking. Because what you're doing is that you have stopped thinking. And then you're just like.

giving AI instructions to do everything for you and that's why you feel left behind because there are other people who are starting to productize this particular thinking. They are doing the thinking. They are actually sitting down and they're putting an effort to build these systems and workflows.

They're building these features and they're selling these features to you. That's what you're paying for. So how do we make money with this?

First, you save money. If you're a creator, you're not paying an editor $500 or more for a video anymore. Second, you can get into editing if you have a visual eye i can coach you anything but i cannot coach you what good is you need to figure that out by yourself and the only way to do this is by making a lot of bad decisions in the beginning and figuring out as you go everything comes down to knowledge and taste and this is exactly how i would do it creators mostly live on youtube and instagram so this is where you you look find 20 creators in your niche who do not have motion graphics yet um take their last video and make three boards from it like make three animations and at this point it really is costing you nothing except your time and you can literally take one of their instagram videos and turn it into something with your system and because you already have a system in place it's going to be so quick for you to produce this it's not going to be the same thing as if you were to do it by hand so now we can mass produce all these templates and all these videos and

even make more boards for you as a creator yourself. Now, most of them are not going to answer. Maybe one or two will.

And if one says yes, well, you have your first client. Now, we can talk more about how to get clients in this space, but that's not what I'm doing today. So I don't want to focus on that too much.

But we need to think about what you're actually selling because it's not simply. designs, you're selling them a system. Their own style, their own templates, their own skill that's built around their channel so that every video now looks like them and every video gets built and done faster so that they have more output.

And obviously right now, quality and quantity matter so much. You only charge once to build the system for them and then you can run this every month. You can do this for short form or for long form.

Short form is where the demand is. Everyone's posting daily, so you can scale it to lots of clients. You're going to charge less, like $50 to maybe $100 for a short.

But long form is where the money is. You can charge something like $300 per video because one good board can be the reason someone actually understands this 20 -minute video. So you help them a lot here.

And look at this, if one long form creator posts once a week, that's about $1 ,200 a month from one client. You can get three of those and that's 3 ,600. You're not putting in a lot of effort, honestly.

And if you're already an editor, it's the same thing, but it's easier because you already have these clients. You just need to add motion graphics on top of every single edit and charge for it. And third, you can build templates like I did, all these boards as HTML files and sell them on Gumroad or StanStore for...

$99 a pop or even build it as a SaaS, like an app. with a subscription plan where people pay every month to use it or download your templates, similar to Katana. And the way that you do that is either you sell the templates or you turn your skill into an MCP.

So instead of you sending people the files, they plug in your skill and all your boards straight into their own cloud. And they pay to use that every single month without really seeing what's behind it. And I'm going to show you how to turn your skills into an MCP and start selling it in the next videos.

Now, I know that I'm making this sound easy, but obviously it's not. Nothing is ever easy. You need people.

you need distribution, you need to build some sort of an audience. But regardless, what you can take from this video is that start selling whatever you know now, even if that's not perfect. And start building that small community, even if it's 10 people.

Because at the end of the day, whether it's a SaaS or workflow and MCP and API, of course, everything comes down to productized knowledge. That's literally how I made $6 million last year. I productized.

what I know. So productize your expertise, my friend, as soon as possible. I actually want you to go in front of the mirror today.

And say three times, productize your expertise, productize your expertise, productize your expertise. See what happens. So close your YouTube tonight, open cloud, take one line from your next script and start making your first board.

Just one, start with one. Was it bad? Was it good?

Turn that into a rule so that AI never does that again. And then make the next one and build your system. Try it at least for a week.

So yes, every board in this video was made this way. It's not instant. Quite literally took me weeks to refine and build 300s of them in my style.

If you want to build systems like this with other people, or get feedback, or just hang out with me and ask your questions, Join Public AI. It's my free community.

And it's for people who want to go further than just making AI content. It's for people who want to build something like a media company, like an app, agents, whatever pushes the boundaries of what you think is possible. We do live calls every week.

So post your boards in there and I'll actually take a look at them. The link is right below. join me today and remember that the animation is is just simply a webpage the skill the workflow the pipeline the thought process behind this whole architecture that's what makes it yours thank you for watching friend thank you for being with me here today and spending your time i hope that you learned something new today and as i always do and say do not forget create without limits this is

The Hook

The bait, then the rug-pull.

Everyone's AI motion graphics are starting to look identical, and Sirio's answer is to build your own system instead of renting one. He gives away 300 boards and the nine steps behind them.

Frameworks