Modern Creator
Danny Why · YouTube

Claude Literally Just Changed YouTube Forever

A walkthrough of the four-prompt Claude Code pipeline the host credits with an unverified 18-million-view month on a channel he never names.

Posted
today
Duration
Format
Tutorial
hype
Views
11.8K
422 likes
Big Idea

The argument in one line.

Claude Code can run an entire AI-video production pipeline, planning, image prompts, art generation, animation, and synced voiceover, from four prewritten prompts chained to Higgsfield and ElevenLabs.

Who This Is For

Read if. Skip if.

READ IF YOU ARE…
  • A creator or marketer curious about using Claude Code as an AI video-production assistant for faceless, animated explainer content.
  • Someone comfortable pasting multi-step prompts into Claude Code and paying per-image AI generation credits through Higgsfield.
  • Anyone looking for a reusable prompt chain that handles planning, image generation, animation, and voiceover in one pass.
SKIP IF…
  • You want proof the growth numbers are real. The host never names, shows, or links the channel he says hit 18 million views, so the claim can't be verified.
  • You're not willing to pay for both Higgsfield credits and the host's $29-a-month private community, which is pitched partway through the video.
  • You want live-action or talking-head footage. This workflow only produces stylized, code-animated motion graphics built from AI-generated images.
TL;DR

The full version, fast.

The host says a new, unnamed YouTube channel went from zero to 18 million views and over 10,000 subscribers in under a month by having Claude Code run the entire production pipeline for short animated explainer videos. The workflow uses four prewritten prompts pasted in sequence: plan the script and shot list with Claude set to Opus and max effort, write detailed per-shot image prompts, generate and background-remove the images through a Higgsfield MCP connection, then animate them with Remotion and sync an ElevenLabs voiceover. A 34-second short cost $1.48 in image credits and took roughly 20 to 30 minutes end to end; the same four-prompt process is shown scaling to a 10-minute long-form video. The channel itself is never shown or linked, so the growth claim can't be checked.

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 – 01:34

01 · The growth claim

The host shows YouTube Studio analytics for an unnamed new channel, claims 18 million views and 10,000-plus subscribers in under a month, credits it to Claude Code and Higgsfield, and promises to teach the process without revealing the channel.

01:34 – 03:20

02 · Wiring up Higgsfield and what it costs

Walks through linking the Higgsfield MCP server to a Claude account from the Higgsfield dashboard's MCP tab, then breaks down the credit pricing: roughly 11 credits per image against plans running from about 800 credits a month up to nearly 10,000 a year.

03:20 – 05:17

03 · Prompt 1: planning the script and shot list

Copies the first prepared prompt from a shared Google Doc, pastes an idea (brain and neurons), and sets the Claude Code model to Opus with effort on Max before sending, so Claude plans the script and shot list rather than jumping straight to animation.

05:17 – 06:11

04 · Prompt 2: writing the image prompts

A second prepared prompt has Claude write its own detailed, per-shot Higgsfield image prompts rather than the creator writing them by hand, on the logic that Claude's prompts are more detailed than most people would bother writing.

06:11 – 07:56

05 · Generating the art

Claude calls Higgsfield to generate and background-remove a set of brain and neuron illustrations; the host returns briefly to the analytics flex before revealing the full image set cost $1.48 in credits.

07:56 – 09:17

06 · Prompt 3: animating and voicing it

A third prepared prompt sends Claude to Remotion to animate the generated images into motion graphics and to ElevenLabs to generate a voiceover, timed to land in sync with the visuals as they appear.

09:17 – 10:12

07 · First result: a 34-second short

Plays the finished animation on camera for the first time, a synced brain and neuron explainer, and reacts that the whole thing, script through voiceover, took roughly 20 to 30 minutes.

10:12 – 12:03

08 · Prompt 4: scaling to long-form, and the pitch

A fourth prepared prompt points the same four-step process at a 10-minute script; before letting it run, the host pitches his $29-a-month private community, then asks Claude to render just the first two minutes so he doesn't have to wait out the full build.

12:03 – 13:06

09 · Reacting to the two-minute cut

Plays back the two-minute result live, narrating the animation as it zooms from the whole brain down to individual atoms with motion-tracked overlays, then closes with another plug for the community.

Atomic Insights

Lines worth screenshotting.

  • A single creator claims a new, unrevealed YouTube channel hit 18 million views and 10,000-plus subscribers in under a month running almost entirely on Claude Code-generated animations.
  • The full video pipeline runs on four prewritten prompts pasted into Claude Code in sequence: plan the script, write image prompts, generate the images, then animate and voice them.
  • Connecting Higgsfield to Claude Code happens through an MCP server, authorized once from the Higgsfield dashboard, after which Claude can call Higgsfield directly to generate images.
  • Claude writes the image-generation prompts itself instead of the creator writing them by hand, on the logic that detailed, consistent prompts produce far better Higgsfield output.
  • Generating the full image set for a 34-second animated short cost $1.48 in Higgsfield credits.
  • Images are generated without backgrounds so Claude can animate each element independently on top of one backdrop, instead of editing flat video clips.
  • The ElevenLabs voiceover is timed to match the animation beat for beat, so the visuals change exactly when the narration names them.
  • Claude Code's model is set to Opus with effort pushed to Max specifically for the planning step, which the host says is what produces a usable shot plan instead of a bad first-pass animation.
  • The same four-prompt workflow that makes a 30-second short is pointed at a 10-minute script to produce a long-form video, generated section by section.
  • The host never names, shows, or links the channel that supposedly earned 18 million views, so the growth numbers driving the whole pitch are unverifiable.
Takeaway

Four chained prompts turn Claude into a video pipeline.

AI VIDEO PIPELINE

A fixed sequence of four prompts, plan the script, write the image prompts, generate the art, then animate and voice it, chains Claude Code, Higgsfield, and ElevenLabs into one explainer-video pipeline.

02Wiring up Higgsfield and what it costs
  • Higgsfield connects to Claude Code through an MCP server, authorized once from the Higgsfield dashboard's MCP tab, after which Claude can call Higgsfield directly inside the same chat.
  • Image generation runs on a credit system, roughly 11 credits per image, with plans ranging from about 800 credits a month to nearly 10,000 a year.
03Prompt 1: planning the script and shot list
  • The first prompt only plans the script and shot list, it never generates the final animation, so Claude is set up on the idea before any rendering starts.
  • Switching the model to Opus and setting effort to Max before sending the planning prompt is presented as the difference between a usable shot plan and a weak first pass.
04Prompt 2: writing the image prompts
  • A second prompt has Claude write its own detailed image-generation prompts rather than the creator writing them by hand, since Claude packs in more descriptive detail than most people bother to write themselves.
05Generating the art
  • Images are generated with their backgrounds removed, which is what lets Claude later animate each element independently on top of one backdrop instead of editing flat clips.
  • The full image set for a 34-second animated short cost $1.48 in Higgsfield credits, offered as proof the pipeline is cheap to run.
06Prompt 3: animating and voicing it
  • A third prompt sends Claude to Remotion to animate the generated images into motion graphics and to ElevenLabs to generate a voiceover.
  • The voiceover is timed to the animation beat for beat, so the visuals change exactly when the narration names the thing on screen.
07First result: a 34-second short
  • The finished 34-second explainer, script through synced voice, took roughly 20 to 30 minutes end to end once the prompts were in place.
08Prompt 4: scaling to long-form, and the pitch
  • A fourth prompt points the same four-step process at a much longer script, generating a 10-minute long-form video with the identical prompt chain.
  • The method is taught for free in the video, but the fuller version, including growth coaching, is sold separately through a $29-a-month private community.
09Reacting to the two-minute cut
  • The host reacts to the AI output live in the video rather than pre-approving it off camera, so the reaction to the motion-tracked zoom and detail reads as genuine rather than scripted praise.
Glossary

Terms worth knowing.

Higgsfield
An AI image and video generation platform that connects to Claude Code through an MCP server, used here to generate background-removed artwork for animated videos.
MCP server
A connector (Model Context Protocol) that lets an AI tool like Claude call an external service, such as Higgsfield, directly from inside a chat session.
Remotion
A code-based video framework Claude Code uses to animate the generated images into motion graphics.
Effort: Max
A Claude Code setting that increases how much reasoning the model applies to a task, used here on the planning prompt to get a usable script and shot plan.
ElevenLabs
A text-to-speech service used to generate the voiceover that Claude times to match the animation.
Resources

Things they pointed at.

08:01toolRemotion (motion-graphics framework Claude uses to animate the images)
07:56toolElevenLabs (voiceover generation)
Quotables

Lines you could clip.

18:55
“It kind of feels like cheating, you know? It feels like we as humans are going to go from manual creativity to literally just speaking our mind to an AI, and the AI will basically do the job for us.”
the clearest statement of the video's actual thesis, AI collapsing the gap between idea and finished video→ 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.

metaphoranalogy
Um, what I'm about to show you is absolutely insane. So I've started a brand new YouTube channel focusing on uploading shorts and let me just show you the analytics. All right, so first of all, I'm not going to reveal the channel yet because I don't want all of you to go and literally steal my niche.
Okay, I am going to reveal the channel but one month later because I'm also making a separate video where I'm showing you how I blow up a channel in 28 days. And as you can see, I'm doing pretty well.
In the last 28 days, I got 18 million views. As you can see over here, we got 300 ,000 views this day, 700 ,000 views this days. And right now we are at almost 3 .5 million views on average, right?
Every single day. And this is in the last literally one week, nine days, something like that. We also got 10 ,000 subscribers over here.
I'm also going to change over here to the last 90 days. So you can see I literally blow up the channel right now. Over here, you can see that it's getting 6 .4 million views every 48 hours, as you can see over here.
If we go to check the live count of the subscribers, we will see that it just started growing right now like crazy. Let's go to the last 90 days, right? So I literally blew up this channel in a week.
I think it's fair to say a week. And I did all of this with the help of Cloud Code and Higgs Field. And in this video, I'm going to show you how you can do the exact same thing.
I'm going to show you the system, the process, and the way I create these videos. The only thing I'm not showing you is the niche, the channel, because as I've said, I don't want you to steal it. Not yet.
I'm going to reveal it a little bit later. So if you want to learn how the process of creating a channel like this one looks like, in this video, I'm going to be teaching you. So...
Let's begin. Now, the first thing we have to do is connect Hixfield to our Cloud Code account. And for that, we're going to be using the MCP server.
There's going to be a link down in the description of this video. If you click on that link, it should send you to the Hixfield dashboard over here. And once you are on the Hixfield dashboard, create an account or log in if you already have an account.
Once you're logged in, go over here to the top left where it says MCP and click on this over here, which says MCP. After you click on this, just select over here where it says Cloud. and then click over here where it says connect Hicksfield.
And this will automatically connect Hicksfield to your cloud account. All you have to do is just authorize and click continue. That's it.
After you do this, you're going to have Hicksfield connected to your cloud code and we can now generate images using cloud code. Now the reason why we want to do this is because if we directly just prompt Hicksfield to generate images for us, our prompts are going to be crappy. But if we connect it to cloud code, Cloud Code is going to write the prompts for us, thus creating way more detailed prompts and thus generating the right images for us.
Now, if you think this is going to be expensive, I promise it's not expensive at all. All we have to do is generate around 5 to 10 images per video. And one image usually costs 11 credits.
And if we check out the pricing over here, you will see that with the basic plan over here, you get almost 10 ,000 credits a year. You can also pay monthly and you will get 800 credits. But this is still enough to make quite a lot of videos.
Now, once this is ready, all you have to do is go to your cloud account. Now, once you are on cloud code, make sure you're on the cloud code section over here at the top left. As you can see right now, we are on the chat.
If we click over here, this is where we will create code. Now, over here is where we will send the first prompt. Now, for this to work, we need to send the right.
prompt and for that I prepared a free Google document with all the prompts I'm going to be using inside this video so you can use them yourself. In the description there's going to be a link which says prompts and if you click on that link it should send you to this Google document over here with all the prompts I'm going to be using and what's cool about these prompts is that they are what I like to call general prompts.
Basically all you have to do is change over here where it says write your idea here. You just change this with your idea and you can basically use this prompt to generate any type of animation for your youtube channel so basically all you have to do over here is copy this entire prompt like this right click copy go back to your cloud code and over here in the chat box just paste the prompt scroll down and where it says write your idea delete this and just type in your idea for this i'm just going to generate something maybe that's like with the brain and neurons i think it looks cool so let me show you how that will look now after you write your idea make sure you change the effort over here of your model first of all make sure you choose opus 5 .5 and once you choose opus 5 .5 make sure you choose the effort to max right, is going to consume a little bit more credits.
We want to use Max for this and you will understand why later on. Now, after you send this prompt, it should take around one or three minutes for Cloud Code to plan your video. So basically what we're doing with this prompt, we're planning the script and we're planning the video.
We're not generating the animation yet. Why? Because if you just generate the animation straight ahead, usually you get a bad result.
So what we're doing, we're preparing, we're training this Cloud Code chat to generate the right animation for us. Now, once Cloud Code is ready, you should have the plan for the video over here for the animation we will create for youtube what you want to do after this is go back to my free google document over here scroll down to step number two and copy this prompt over here right click and copy my prompt then go back to clock code and over here paste this text paste this prompt and send the message so what we have now is basically the script and the plan for the animation.
And now the second prompt is asking Cloud Code to write the prompts for the animation. So we're not writing prompts to generate these images. Cloud Code is going to write the prompts for us using the second prompt over here.
Now, the reason why you want to do this is because you're probably terrible at writing prompts, right? But Cloud Code can actually write a huge prompt full of details so that the image looks exactly as it should for the script for the animation. right?
And you will see what I'm talking about once I'm done creating the animation. Your mind will be blown away by the animations and edits Cloud Code is capable of doing when you connect it to Higgs field. And I think it should also show us over here how much we're paying.
It's going to show us how many credits we've spent to create these images and how much it costs in dollars. And I'm telling you, it's not as expensive as you think. I mean, considering the fact that you can go viral like I just did over here, let me bring back my analytics.
I mean, this is just insane. 6 .5 million views in the last 48 hours, right? We're getting almost 200 ,000 views per hour.
If I click over here in the last 365 days, we will see the channel blew up literally. last week. You can see over here, right?
So as you can see, Cloud Code is using Higgs field to generate the images. And what's cool is that they don't have backgrounds because you can remove the background using Higgs field. So that is why we're also using Cloud Code to do those things for us because we don't have to manually do them ourselves.
And now because they don't have a background, Cloud Code can take them and animate them for us. And you will see what I mean. It's going to blow your mind how good this looks.
I'm a video editor myself. I've been editing videos for more than 20 years. I use After Effects, Premiere Pro, CapCut, Sony Vegas.
I've used a lot of programs. And I'm barely touching them anymore because Cloud Code does a much better job at editing for me. Now, once Cloud Code is done creating these images using Higgs field, as I've said, it's going to show us how much these images over here cost.
So over here, as you can see, The total cost was $1 .48. That's how much I had to pay to create these images.
Now, this is where the magic is going to happen. Go back to my Google document over here, scroll down to step number three, and copy this entire prompt, this entire text. Right -click, copy, go back to Cloud Code, and over here, paste the prompt.
And this prompt will ask Cloud Code... to start creating the animation for us because we now have the plan the script the images cloud code will now use remotion to create a beautiful motion graphic animation with these images over here talking about the brain and how the brain works This usually takes around 5 minutes, 10 minutes, 15 minutes.
It really depends. Now also, just so you know, I also have Eleven Labs connected to Cloud Code. So it's also going to generate the voiceover for me using Eleven Labs.
So basically once this animation is ready, once the animation is done, I'm going to have an animation with a voiceover that's already edited and has motion graphics. So I'm just going to wait for Cloud Code to create the animation and I'm going to show it to you right away. So as you can see over here, Cloud Code used Eleven Labs to generate the voiceover.
if I hit play it should sound like this. This is a human brain and the cells inside it never stop talking. Let me increase the volume.
A neuron. Its branches catch signals. And the crazy part about all of this is that the voiceover is going to be synced to the visuals.
Whatever the voiceover is saying the visuals are going to be shown on the screen. So if the voiceover says neurons Cloud Code is going to show neurons on the screen.
It's going to be synced for you. It's going to be edited for you, which is kind of mind -blowing, and it kind of feels like cheating, you know? It feels like we as humans are going to go from manual creativity to literally just speaking our mind to an AI, and the AI will basically do the job for us.
And we just speak and speak and speak to the point where we create something awesome. It's kind of weird. All right, cloud code is ready and we have the animation over here.
I'm going to put my headphones so I can also listen to the voiceover. And as you can see, you actually have the video on your computer. So I'm just going to open this and play it.
34 seconds. So this is going to be a YouTube short, basically. This is a human brain and the cells inside it never stop talking.
Here's one of them, a neuron. Its branches catch signals from other neurons. When enough arrive, it fires.
An electric pulse runs down this long tail. At the end, there's a gap. So the cell sends chemicals across.
The next neuron picks them up, and the message keeps going. You have about 86 billion neurons, and this is happening in your head right now. This is insane.
Like what? This took 20 minutes, 30 minutes to create? And trust me.
It doesn't end here because if you go back to my document and you scroll down at step number four, if you copy this prompt, what this prompt does is basically going to create an entire 10 minute animation for you for a long form video. So I'm just going to paste this over here and start generating. I'm going to ask Cloud Code to stop at around two minutes.
So I'm just going to wait until Cloud Code actually starts creating the animation. Then I'm going to stop it and then I'm going to show you the video. before i show you the final video if you guys want to learn more about youtube and how to grow on youtube i have a private community that costs 29 a month to join you can join as a free trial inside i have courses i help my members grow so over here if i go to the community you'll see that i help my members grow over here he got monetized in three weeks there's going to be a link down in the description where you can check out my private community but as i said i don't want to wait so i'm just going to ask cloud code to send me the first two minutes first so Hey, can you first send me the first two minutes because I don't want to wait one or two hours for you to make the full animation.
I want to see how it looks in the first two minutes. Now, as you can see over here, Cloud Code is generating all the images we need for this animation using Higgs field. And as you can see, there are a lot of animations over here.
It's basically doing the entire job for us, which is really cool. All right, Cloud Code has finished rendering the first two minutes of the video. If you're ready to have your mind blown away, get ready because The way Clot Code is able to edit is absolutely insane.
I'm going to react to this for the first time as well. I'm going to actually play the MP4 video because it will have a higher resolution. And by the way, this took in total around 30 minutes.
So let's react to it. Right now, the signals in your brain are being carried by sodium. The same sodium that's in table salt.
Every time you see something... This already blows my mind because it has motion tracking. Remember something.
or decide anything tiny charged atoms of it rush into your brain cells so in this video so you see what cloud code did it used the images created with hicks field and then it animated code on top of it so we have code and images and it creates a really interesting video i want to start with the whole brain and zoom in step by step until we get to those atoms and by the end you'll know what a thought is actually made of Okay, let's start with the whole thing.
This is a human brain. An adult one weighs about 1 .4 kilograms. This is insane.
And about three -quarters of it is water, 2 % of your body weight. You can see it added this motion graphic over here, so it added this motion graphic over here. It looks pretty cool.
But it uses about 20 % of your energy. That's about 20 watts. Yeah, you get the point.
I don't know what to say. If you want to learn more, there's going to be a link down in the description where you can join my private community. And I teach things like this inside my private community.
Peace.
The Hook

The bait, then the rug-pull.

The video opens on an unverified flex: a brand-new YouTube channel, never shown or named, that the host says pulled 18 million views and more than 10,000 subscribers in under a month using nothing but Claude Code and an AI image tool called Higgsfield to generate animated explainer shorts.

Frameworks

Named ideas worth stealing.

01:35list

The 4-Prompt Video Pipeline

  1. Plan the script and shot list (Claude Opus, effort Max)
  2. Write the per-shot image-generation prompts
  3. Generate and background-remove the images via Higgsfield
  4. Animate with Remotion and voice with ElevenLabs, synced to the script

Four prewritten prompts, pasted into Claude Code in a fixed order, chain Higgsfield image generation and an ElevenLabs voiceover into one finished, synced explainer video.

Steal forany automated faceless-video or AI-explainer content pipeline
CTA Breakdown

How they asked for the click.

VERBAL ASK
10:40product
“If you guys want to learn more about YouTube and how to grow on YouTube, I have a private community that costs 29 a month to join. You can join as a free trial, inside I have courses, I help my members grow.”

Dropped as a mid-roll soft-sell right before the biggest technical payoff (the long-form generation), then repeated almost verbatim as the video's closing line, a low-pressure plug rather than a hard close.

FROM THE DESCRIPTION
PRIMARY CTAWhere the creator wants you to go next.
OTHER LINKSAlso linked in the description.
Storyboard

Visual structure at a glance.

cold open
hookcold open00:00
analytics flex
promiseanalytics flex00:44
connecting Higgsfield
valueconnecting Higgsfield02:12
pasting prompt 1
valuepasting prompt 104:10
image cost reveal
valueimage cost reveal07:07
finished short plays
valuefinished short plays09:46
community pitch
ctacommunity pitch10:40
closing pitch
ctaclosing pitch13:06
Frame Gallery

Visual moments.

One-click upgrade to your Google

Get more breakdowns in your search results

Add Modern Creator as a preferred source and Google shows you more of our breakdowns in Search, Top Stories, and AI Overviews. It only changes what you see, and you can undo it in your Google settings anytime.

Add to Preferred SourcesOpens your Google source preferences with us pre-loaded. Tick the box and you're done.
Watch next

More from this channel + related breakdowns.

29:59
Danny Why · Tutorial

Claude Code Just Changed CapCut Forever

A screen-recorded build of a full 3D motion graphic inside CapCut, using a Claude Code-generated asset and an image-to-video AI to carry it past the timeline's edge.

August 21st
13:13
Danny Why · Listicle

Claude Code Just Changed The Internet Forever

A weekly AI news roundup: a skateboard mod spliced into World of Warcraft, Anthropic's new Claude Motion, a free seven-app Adobe alternative built with Opus 5.5, and a Claude agent that almost submitted a false tip to a real police department.

October 12th