Big Idea

The argument in one line.

Code-driven animation tools have made faceless video production nearly free and equally available to everyone, so the tool itself stops being a competitive edge and a creator's own judgment and voice become the only thing left to differentiate a channel.

Who This Is For

Read if. Skip if.

READ IF YOU ARE…

You want to build a faceless YouTube channel and are deciding between AI slideshow tools and something more visually distinctive.
You already use Claude Code or a similar AI coding tool and want a concrete, code-based use case beyond writing software.
You're evaluating whether to use AI voice narration or record your own, and want the actual tradeoff spelled out.
You want to understand how AI-generated video actually gets assembled (layers and code) rather than treating it as a black box.

SKIP IF…

You want a video-editing tutorial for traditional camera-shot content — this is entirely code-generated, faceless production.
You're looking for a no-setup, browser-only tool — this workflow requires installing VS Code and Claude Code and editing project files.

TL;DR

The full version, fast.

One open-source Claude Code project generates three unrelated video styles — an animated chess lesson, an illustrated kids' story, a paper-cutout documentary — by stacking layers (background, characters, labels, camera moves) on a timeline with the Remotion animation engine, similar to Photoshop but with moving layers. Claude acts as director: it plans scenes, writes the animation code, adds AI voice and word-synced captions, and follows a style guide. Getting started takes four steps — install VS Code and Claude Code, clone the free GitHub project, add API keys, and describe the video you want in plain language. The catch: because the tool is now free and available to everyone, output volume stops being an advantage, and creators should record their own voice instead of the default AI voice, since a recognizable voice — not raw video output — is what builds an audience that recognizes and trusts the creator later.

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 – 01:31

01 · Claude makes any faceless video

Cold open teasing three unrelated video styles built from one project, framing the video's three-part promise: watch the demos, learn how it works, then build your own.

01:31 – 03:30

02 · The three videos in full

Each demo plays in full — an animated chess tutorial with arrows and captions, a soft illustrated kids' story, and a paper-cutout Vox-style coffee documentary — same project, three different styles, with a nod to 3Blue1Brown-style code-animated education as a proven format.

03:30 – 06:01

03 · How it actually works: layers, not magic

Debunks the 'type a prompt, get a finished video' assumption. Shows the coffee documentary broken into stacked layers (background, map, labels) composited like Photoshop but moving, explains Remotion as the animation engine and the AI voice/caption-sync layer, and frames Claude as the director who plans scenes and writes the animation code.

06:01 – 07:59

04 · Build your first video: VS Code + Claude Code, 4 steps

Walks through installing VS Code and Claude Code, then the four steps to a first video — clone the GitHub project, open it in VS Code, add API keys to .env, and tell Claude Code what you want. Emphasizes that Claude writes a full plan (hook, scenes, timing) before building, which the creator reviews first.

07:59 – 09:12

05 · The faceless trap: why your own voice still wins

Argues that once everyone can produce faceless video, output stops being an advantage — the creator's judgment and voice are what's left. Makes the case for recording your own voice instead of the default AI voice, citing his own accented, sometimes-mocked voice as what actually grew his channel to a million subscribers.

Atomic Insights

Lines worth screenshotting.

AI-generated faceless video is closer to editing than to one-shot generation: it's separate layers (background, characters, labels, camera) composited on a timeline, the same way Photoshop stacks layers, except the layers move.
One project and one AI coding tool can output three completely different visual styles — chess tutorial, illustrated bedtime story, paper-cutout documentary — because each style is just a different layer set, not a different tool.
Code-animated educational content (the 3Blue1Brown-style format) is proven, high-value YouTube content, not a gimmick — it used to require real animation skill that AI-assisted code generation now largely removes as a barrier.
The animation engine behind the demoed videos is Remotion, and the AI voice is generated with a third-party voice service that syncs captions to the exact spoken word.
The actual workflow is four steps: install VS Code and Claude Code, clone a prebuilt open-source project from GitHub, add API keys for voice and image generation to a .env file, then describe the video in plain language.
Before Claude Code builds anything, it writes out a full plan first — hook, scenes, timing — so the creator can review and correct it before rendering starts.
Once a production barrier drops to near-zero for everyone, raw output volume stops being a competitive edge; the channels that grew big did it with one good video at a time, not by posting fifty videos a week.
Default AI voice narration is the fastest way to publish a video, but it makes the output indistinguishable from anyone else running the same pipeline — nothing about it is recognizably the creator's.
A creator's own voice, accent and imperfections included, is what turns a one-time viewer into someone who recognizes the creator later — including if they eventually step in front of a camera.
An AI voice can earn views on a single video; a creator's own voice is what gets the creator actually known and remembered across videos.

Takeaway

The tool got cheap; your voice didn't

WHAT TO LEARN

Code-driven, layered animation tools have made faceless video production nearly free for everyone, so the only thing left to differentiate a channel is judgment and a recognizable human voice.

AI-generated video is closer to editing than to one-shot generation: an animation engine composites separate layers (background, map, character, captions) on a timeline the way Photoshop stacks layers, then moves a camera across them.
A single project can output wildly different visual styles (chess tutorial, illustrated children's story, paper-cutout documentary) because each style is just a different layer set and prompt, not a different tool.
Long-form educational animation made with code is a proven, high-value YouTube category, not a gimmick — it used to require real animation skill that AI-driven code generation now removes as a barrier.
Once a production barrier drops to near-zero for everyone, output volume stops being a competitive edge; the video's chosen subject and point of view becomes the only thing left to differentiate a channel.
Default AI voice narration is the fastest way to publish, but it produces content indistinguishable from anyone else's identical pipeline — a viewer can watch it without ever learning who made it.
A creator's actual voice, imperfections included, is what converts a one-time viewer into someone who recognizes and trusts the creator later, including if they eventually appear on camera.
The path into this workflow is genuinely low-friction: install VS Code and an AI coding tool, clone a prebuilt open-source project, add API keys for voice and image generation, and describe the video in plain language before any code gets written.

Glossary

Terms worth knowing.

Remotion: A code-based framework for building and rendering video programmatically, letting a developer control every frame and layer through code instead of manual timeline editing.
Faceless YouTube: A channel format where videos are produced and narrated without the creator appearing on camera, often relying on AI voice, animation, or stock footage instead.
Claude Code: An AI coding tool that runs inside a code editor and can plan, write, and execute code changes from plain-language instructions.
Word-synced captions: On-screen subtitles that highlight or reveal in time with the exact word being spoken in the AI-generated voice track, rather than appearing line by line.

Resources

Things they pointed at.

04:18toolRemotion

04:27toolElevenLabs (AI voice)

02:41channel3Blue1Brown

05:47toolclaude-faceless-shorts-creator (GitHub project) ↗

06:24linkFree beginner guide (learnwithhasan.com) ↗

06:44productBuild With AI course ↗

Quotables

Lines you could clip.

03:54

“It's like Photoshop, but the layers move.”

tight one-line explanation of the whole pipeline→ TikTok hook↗ Tweet quote

07:23

“The tool is not the advantage anymore. Everyone has the tool, and it's almost free or at least cheap. The only thing left that really matters is what you choose to make, and that it's actually you.”

the video's clearest thesis statement, self-contained→ newsletter pull-quote↗ Tweet quote

08:13

“But that same voice and that same mocking helped me grow on YouTube and reach 1,000,000 subscribers.”

personal proof point that lands right before the payoff line→ IG reel cold open↗ Tweet quote

08:29

“An AI voice may get you views. Your voice gets you known.”

punchy contrast line, doubles as the video's title card copy→ TikTok hook↗ 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.

metaphoranalogystory

Look at this. A chess lesson short video. Pawn to e four.

Bishop c four. This one is a kid's bedtime story.

Today, Pip is chasing a butterfly.

This is a Vox style documentary about coffee.

Rulers panicked. Mecca bandits. Three videos,

no camera, no video editing, and no stock footage. Now when people hear faceless AI videos, they picture one of two things.

Either those weird AI clips with melting hands or a lazy slideshow with robotic voice just to farm views. But what I noticed is that everyone is fighting over the wrong thing. They are fighting over the tool, and the tool became now the easiest part as you will see in this video.

Creating a video like this one is so easy now. Even a 70 year old kid could describe it and Claude can build it in minutes. So in this video, I'm going to show you three things.

One, these three videos fast so you can see the range of what we can build with your own eyes. Then how it actually works, how to create such videos. And it's not, hey, Claude, make me a video.

It's something way better and different. And then how you can make your first video today. And, yes, you'll get this project for free by the end of this video.

So let's start with the three videos. Video one, the chess tutorial. F seven, the weakest square in black's camp.

Queen h five. If black plays the natural move Every piece, every arrow, every caption, that's not a screen recording of a chest application.

It's all drawn and animated in code with Claude. Look how clean the moves are.

Video two, little pip, the kid's story. Far and farther,

deep into the trees. Completely different feeling.

Soft, warm, illustrated. The art is AI generated, the same engine animates it and reads it out loud.

And video three,

coffee. It conquered the world. Legend says an Ethiopian goatherd.

This one looks like a mini vox style documentary. Paper cutouts, a real map, a camera flying across it, like something of a channel with whole motion team behind it.

And here's what I want you to understand. These are three different tools. It's one, same project, same setup, three completely different styles.

And there are 12 more already in this project, math, security explainer, physics, music theory, and much more.

And here's the thing about this style. It's not gimmick. It's some of the most valuable content on YouTube.

You know three blue one brown channel? Millions of people learn math from his animations. And the interesting part, he makes those animations with code.

The exact same idea you just saw. Or BraitheMathGuy or CodeCloud, this style, clean, animated, teaching one thing well is proven, and channels get huge doing it.

The only difference that this used to take a lot of time and requires a lot of skills. Now Clothe can do all of it in minutes. But how?

How does one tool, one project make a chess animation video and a paper documentary?

Here's the mistake everyone makes. They think you can type, make me a chess video, and AI spits out the finished video.

Maybe in the future this happens, but it's not now. What's actually happening is way closer to editing than generating.

Let me explain. Take the coffee vox style documentary I showed you. On screen, it looks like one video, but it's not.

Watch this.

It is layers stacked on top of each other. The paper background is one layer.

The map is another. The little cup, their outline, the labels, and then a camera moves across all of them.

It's like Photoshop, but the layers move. Across the Red Sea to Yemen, where Sufi monks brewed them to The chess video is the same idea.

The board is a layer. The pieces are layers. The arrows, the captions are layers too.

All animated with code. And the engine powering all this underneath is called Remotion.

Maybe you heard about it lately. It was trending. Then we have the voice.

The AI read the script loud. And this project I built uses 11, and it aligns the captions up to the exact word as it's spoken.

Now one thing I want to be honest with here, that voice you just heard, it is an AI voice. By default, every one of these videos I created for this demo is narrated by AI. Hold on to that because at the end, I'm going to tell you why I think you should not use AI voice and what to do instead.

And it's really important, and maybe it's what will make you succeed on YouTube or any other platform if you wanna create short videos. Okay. So where does Claude come in?

Think of Claude as a director. You describe the video you want. Claude plans it out, the scenes, the timing.

It writes the code for the animation. It adds the voice and the captions. It even designs sound effects, and it follows a style guide, so everything stays on brand.

Okay. So it's really powerful, but is it actually easy to start? Let me show you how fast you can create your first video with Cload.

So you need two things installed, Versus Code and Cload Code. Both are free to install.

And if you don't know what those are, Versus Code is just a program you open the project in. And Clode Code is simply Clode running inside it and allows you to build things with AI. I think you know this.

That's it. Then we have four steps. One, download the project I built for you from GitHub in one command or one click.

Two, open it in Versus code. Three, add your keys.

There's a file called dot ENV, and you paste in your keys for the voice and the images. Number four, open Cload code and tell it what you want.

Let's say you want a short video about the water cycle for kids. You just type it and why Clothe is planning. This is the part I want you to see.

It doesn't just run randomly. It writes out the whole plan first, the hook, the scenes, the timing.

You read it. You fix what you don't like, and then it builds. That's your first short.

The sun heats the ocean. Water rises into clouds, then falls back down as rain. If you want the real step by step guide, every key, where to get it, how to plan a view that's actually good, I made a full guide on my website, and, yes, it's free, and it's linked below.

And if you wanna go further and learn to build systems and apps with AI like this yourself, you can check my full course on building with AI, which I think is the most important skill in 2026 and beyond. It's also linked below.

So here's the thing. You can set up and create your first video actually in minutes. The barrier is basically gone now, and that is the part I actually want to talk about because it really changes how you think and what really matters.

So as we mentioned, everyone can now create a video, even a kid, And that sounds like a great idea, but it's also a trap. Because when everyone can produce videos in hundreds, then videos is worth nothing.

The tool is not the advantage anymore. Everyone has the tool, and it's almost free or at least cheap. The only thing left that really matters is what you choose to make, and that it's actually you.

Look at those channels again. Three blue one brown, bright the math guy, cold cloud, and many others. None of them blew up because they posted 50 videos a week.

They blew up because every single video was worth watching, which brings me back to that AI voice. The project I built uses it by default because it's easy, and I wanted to show you that you can build anything with AI. But if you are serious, use your own voice even if it's not perfect like me.

And, yes, many mop my accent in the comments. But here I am. I didn't stop.

I have been on YouTube for more than ten years now. And yes, some comments may hurt me. I'm a human.

But that same voice and that same mocking helped me grow on YouTube and reach 1,000,000 subscribers. Your voice is how people start to know you, to trust you, to recognize you.

An AI voice may get you views. Your voice gets you known.

And maybe one day you decide to step in front of the camera, and when you do, people already know your voice. You are not a stranger. You are someone they have been listening to for months.

Think about it. So in short, create videos that's actually worth someone's time.

That's the whole point. So AI removed the excuse, but it didn't remove the point.

The project is free. The guide is free. Everything is linked below, and it's open source.

I keep updating on GitHub. More styles, more effects, more of everything.

So don't forget to give it a star on GitHub. And if you learned something new today, smash the like button, and see you in the upcoming video.

The Hook

The bait, then the rug-pull.

Three completely different videos — a chess tutorial, an illustrated bedtime story, a paper-cutout documentary — no camera, no editor, no stock footage, all from one project. Hasan opens by playing all three back to back, then pulls back the curtain: it's not a magic "make me a video" prompt, it's layered, code-driven scenes with Claude directing the build.

Frameworks

Named ideas worth stealing.

03:54concept

Layers, Not Generation

The finished video isn't one generated clip — it's separate stacked layers (background, map, character art, labels, captions) on a timeline, composited like Photoshop, except the layers move and a virtual camera pans across them.

Steal forexplaining any AI-assisted production pipeline to a non-technical audience

05:44list

4 Steps to Your First AI Video

Download the open-source project from GitHub
Open it in VS Code
Add your API keys to a .env file
Tell Claude Code what video you want

The full on-ramp into the workflow, positioned as low-friction because both required tools (VS Code, Claude Code) are free to install.

Steal forany code-generated content pipeline onboarding

CTA Breakdown

How they asked for the click.

VERBAL ASK

06:44product

“if you wanna go further and learn to build systems and apps with AI like this yourself, you can check my full course on building with AI”

Soft pitch, buried mid-video right after handing over the free guide and free open-source project — low-pressure, framed as an optional next step rather than a hard sell.

MENTIONED ON CAMERA

05:47toolclaude-faceless-shorts-creator (GitHub project) ↗

06:24linkFree beginner guide (learnwithhasan.com) ↗

06:44productBuild With AI course ↗

FROM THE DESCRIPTION

PRIMARY CTAWhere the creator wants you to go next.