Modern Creator
Ryan | AI Content Automations · YouTube

How I Fully Automated Video Editing With Claude Code and Remotion

A YouTuber hands Claude Code a boring talking-head recording and gets back motion graphics, camera zooms, and a pitch for a $2,500-per-client AI editing service.

Posted
1 months ago
Duration
Format
Tutorial
educational
Views
12.3K
260 likes
Big Idea

The argument in one line.

Claude Code can direct Remotion and WhisperX to turn a raw talking-head recording into a fully motion-graphic-edited video, and the same automated pipeline can be resold as a niche AI video editing service.

Who This Is For

Read if. Skip if.

READ IF YOU ARE…
  • A solo creator doing talking-head YouTube videos who wants motion graphics and camera zooms without manual editing work.
  • Someone weighing an AI-editing side hustle who's comfortable prompting Claude Code, Codex, or a similar coding tool rather than using a drag-and-drop editor.
  • An agency or freelancer looking to build a productized AI editing offer niched to one type of business instead of competing on hourly rates.
SKIP IF…
  • You want a no-code, click-and-drag video editor. This pipeline runs entirely through prompting Claude Code or Codex and reading Remotion/React output.
  • You're looking for a free plug-and-play tool. Building it yourself takes real setup time, and the polished version is sold inside a paid Skool community.
TL;DR

The full version, fast.

Ryan builds his videos with an AI pipeline: Claude Code or Codex directs Remotion, a React-based rendering engine, while WhisperX supplies word-level timestamps so animations land exactly on the spoken words. The core technique is a three-layer Vox-style prompt (background, midground, foreground) for layered 3D scenes, plus a second method of feeding Claude a reference video's link and asking it to study and rebuild that style as a reusable template. He argues every video made this way doubles as proof of service quality, and the real money isn't in hourly rates but in niching the offer toward one type of business and charging a flat monthly retainer.

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.

00:0000:39

01 · The video is the demo

Ryan opens by revealing every effect in the video was made by Claude Code and Remotion, then promises to show how to sell the same pipeline as a $2,500-per-client AI editing service.

00:3901:39

02 · The three-tool stack

Claude Code (or Codex) directs the project, Remotion renders motion graphics in React like a programmable CapCut, and WhisperX stamps word-level timestamps.

01:3902:33

03 · Why word-level timestamps matter

WhisperX's timestamps let Claude sync animated graphics to the exact spoken word, which is presented as the mechanism that builds viewer trust.

02:3303:29

04 · The three-layer Vox-style prompt

Directing specific movement beats letting Claude wing it; a background/midground/foreground prompt structure produces the layered 3D animated scene.

03:2904:41

05 · Reverse-engineering a style from YouTube

Feed a reference video's link to Claude or Codex, have it study the animation frame-by-frame, and save the recreation as a reusable template (the Chris Moran example).

04:4105:46

06 · The pitch: join the Skool engine

Ryan offers his prebuilt engine and weekly training calls inside his Skool community, including newly added Shorts/Reels training.

05:4606:38

07 · Ryan's track record

Clients have earned a collective $2M+; Ryan's own $2,500 automation offer scaled to $200K in about 90 days back in 2024.

06:3808:23

08 · Sell the niche, not the hour

Every video doubles as a demo of the service; generalized AI editing competes on Fiverr/Upwork hourly rates, so niching to one vertical (real estate, roofers, plumbers, HVAC) supports $2,500/month pricing instead.

08:2309:15

09 · Pricing structure and close

A one-time install fee plus a monthly retainer, or a discounted 3-month package at $1,997, then repo links for viewers who want to build it themselves.

Atomic Insights

Lines worth screenshotting.

  • Word-level timestamps from WhisperX let Claude sync motion graphics to the exact spoken word, which is what makes automated editing look intentional instead of random.
  • A three-layer prompt, background, midground, and foreground, is what produces a 3D-feeling Vox-style animated scene instead of a flat one.
  • Letting an AI editor fully improvise the visuals produces inconsistent results; directing specific movement, like which way to glide or what shape appears, is what makes the output usable.
  • You can clone someone else's editing style without hiring an editor by feeding the reference video's link to an AI coding tool and asking it to study the pattern frame-by-frame.
  • Once an AI-recreated animation style is dialed in, saving it as a reusable template turns a one-off edit into a repeatable production system.
  • Publishing videos that show off an automated editing style turns every upload into a demo reel for anyone selling that editing as a service.
  • Generalized AI video editing competes on Fiverr and Upwork rates near $50 an hour; a niche-specific offer for one type of business can charge $2,500 a month instead.
  • Business owners are framed as equating a higher price with quality and a lower price with cheapness, which is the argument for a narrow, expensive offer over a broad, cheap one.
  • A three-month package discounted to $1,997 is pitched as a way to trade a lower headline price for more stable recurring income.
  • Five clients on a discounted three-month retainer model was framed as clearing more than $120,000 a year.
Takeaway

Niche the service, don't rent out the hour.

WHAT TO LEARN

An AI editing pipeline built from Claude Code, Remotion, and WhisperX only becomes a business once it's aimed at one specific type of client instead of sold as a generic hourly service.

02The three-tool stack
  • The pipeline runs on three connected tools: Claude Code or Codex as the director, Remotion for rendering motion graphics in React, and WhisperX for word-level audio timestamps.
  • Remotion works like a programmable version of CapCut: instead of dragging clips on a timeline, Claude writes code that produces animations, captions, layouts, and effects, then Remotion renders the final MP4.
03Why word-level timestamps matter
  • WhisperX's word-level timestamps give Claude a timeline to work from, so an animated graphic can pop in at the exact word it illustrates instead of drifting off-sync.
  • Matching what's said to what's shown on screen at the word level is presented as the mechanism that builds viewer trust in a talking-head video.
04The three-layer Vox-style prompt
  • Letting Claude improvise the visuals produces inconsistent results, so directing specific choices, like which way an avatar glides or what shape appears, is what makes AI-generated motion graphics usable.
  • A reusable three-layer prompt structure, background, midground, and foreground, is what produces a layered, 3D-feeling animated scene instead of a flat one.
05Reverse-engineering a style from YouTube
  • A reference video's link can be fed directly to Claude or Codex with instructions to study it frame-by-frame and recreate the animation style in Remotion.
  • Once an AI-recreated style is close enough, saving it as a template turns a one-off recreation into a reusable asset for future videos.
06The pitch: join the Skool engine
  • The pitch for the paid community is woven directly into the tutorial narrative right after a genuinely useful technique is demonstrated, rather than opening the video with the sales pitch.
07Ryan's track record
  • The credibility claim, $2M collective client earnings and a $2,500 automation offer scaled to $200K in about 90 days back in 2024, is placed after the free value, not before it.
08Sell the niche, not the hour
  • Every video made with this visible editing style is framed as a demo reel: the motion graphics on screen double as proof of the service being sold.
  • Generalized AI editing competes on Fiverr and Upwork hourly rates near $50; niching to one vertical, like luxury real estate, is pitched as a way to charge $2,500 a month instead.
  • Business owners are framed as valuing craft over price, so an expensive, narrow offer is argued to outsell a cheap, broad one.
09Pricing structure and close
  • A one-time installation fee plus an ongoing monthly retainer, or a discounted three-month package at $1,997, is offered as a way to convert one-off setup work into recurring income.
  • The repo links for Remotion and WhisperX are given directly so a viewer can start building the same pipeline without joining the paid community.
Glossary

Terms worth knowing.

Claude Code
An AI coding agent that writes and runs code from natural-language instructions. Here it acts as the 'director' that writes the Remotion code for each edit.
Remotion
An open-source framework for building and rendering video in code using React, letting an AI produce motion graphics, captions, and transitions as programmatic components instead of manual timeline edits.
WhisperX
An open-source speech-to-text tool that returns word-level timestamps, giving Claude a precise timeline of when each word is spoken so animations can sync to it.
Vox-style animation
A layered, illustrated motion-graphics look popularized by Vox's YouTube videos, built from background, midground, and foreground layers moving at different depths.
Codex
OpenAI's AI coding agent, used interchangeably with Claude Code in this video as the tool that writes the Remotion code.
Resources

Things they pointed at.

00:57toolRemotion
01:28toolWhisperX
04:06channelChris Moran (YouTube)
Quotables

Lines you could clip.

00:00
Everything you're seeing right now from the motion graphics, the transitions, the animations was all done with Claude code and Remotion.
cold-open reveal that proves the video's whole premise in one lineTikTok hook↗ Tweet quote
01:10
Claude Code is a master editor who directs the whole project
tight, quotable framing of the tool's roleIG reel cold open↗ Tweet quote
02:43
Your creativity is way more powerful than AI.
short, contrarian pull-quote inside an AI tutorialnewsletter pull-quote↗ Tweet quote
06:43
This means every video you create is a demo of your services.
sums up the core business insight in one sentencenewsletter pull-quote↗ Tweet quote
07:27
People who buy hourly rates always look for the cheapest possible person, but real business owners care less about price and more about craft.
the pricing-philosophy punchline of the whole pitchTikTok 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
Everything you're seeing right now from the motion graphics, the transitions, the animations was all done with Claude code and Remotion.
I gave Claude a boring talking head video with dead air, misspoken sentences, and zero edits. Then it cleaned the audio, added pro level transitions, and overlaid my screen with premium motion graphics.
But Opus five over delivered by adding b roll, Vox style animations, sound effects, and camera zooms. So in this video, I'm gonna show you exactly how to do it.
And then I'm gonna show you how you can build an AI side hustle that earns you $2,500 per client just by editing their videos with AI. So So here's exactly how to do it, my friends.
You're really only connecting three pieces of software as the foundation, but it's not just plugging them into each other and then boom, you have an AI editor. Although it kinda is. You need clot code or codecs if you prefer.
That's fine too. Remotion and WhisperX.
And as you're seeing on my screen, it's just a project folder that lives in my documents with some Python programming that tells each program how to speak to each other. Claude Code is a master editor who directs the whole project, and Claude uses Remotion to create all of the motion graphics, timing, and transitions just like you'd see with traditional editing software.
Remotion is basically a programmable video editor that uses React, which is the same technology used to build websites. But instead of manually dragging clips around, Claude can create animations, captions, layouts, and effects with code.
Then Remotion renders everything into real m p four videos. So Claude is using Remotion in the same way you would use CapCut or another video editing software. Then there's WhisperX, which is your audio editor.
It basically stamps your words so it can give Claude a proper timeline to follow when adding the motion graphics in Remotion. For example, you're seeing an animated graph on my screen right now about uploading three tutorial videos per week to grow YouTube viewership. That animated graph is showing because Whisper X tells Claude what is being said and when.
So this lets Claude create an animation that goes with the spoken words, and that's how you build trust with viewers. You take what is said and match it with what is seen. And all of that is pretty simple, but the part that takes some setting up is curating and editing templates and animations.
For example, look at how I glide in a certain direction when I'm making a point. You could wing it and just tell Claude to create the entire video for you, but I don't recommend it. It's worth you taking the time to instruct it.
So when I was first programming my engine, I would tell Claude to design a way for my avatar to gently glide to the left. Then I tell Claude to make sure there's enough space so we can include a flowchart, animated graph, or a box style animation.
But every time I let Claude just wing it for me, I got mixed results. Some were good, but your creativity is way more powerful than AI. So here's exactly how to create the Vox style layering animations with your own taste.
You're prompting Claude to keep three things in mind. A background image, a set of mid ground images, and a set of foreground images. This is what gives that layered three d effect that we all love.
For example, to recreate something similar to the opening layered animation where I mentioned motion graphics, the transitions, and the animations, you tell Claude something like this.
Create a three d layered Vox inspired animated scene. The background should have a rotating sun with sun rays and clouds and birds that fly onto the screen from the left and the right. The mid ground should have three animations that match my script and are perfectly timed to the spoken words of motion graphics, transitions, and animations.
The foreground imagery should feel like an outside workstation in a minimalist tropical setting. Sometimes Claude or Codex will nail it and you'll have no changes. And other times, it gets close and you just need to make a few tweaks.
Now the other way to get stylistic animations is to search YouTube for a style you like, then get the link to that video and feed it to Claude. Then tell Claude to study the video frame by frame and recreate the animation style from the motion graphics to the timing. It will get you a draft that you can update and modify for your engine.
So here's an example of how that works. I follow a guy on YouTube named Chris Moran, and he does an excellent job of breaking down how to edit videos in the Vox style. He does it with Adobe, but I want AI to do it.
So I found this short where he broke down a Vox edit that has a short three point intro card. Then I took the link to that short, put it into Codex, and asked Codex to recreate it in Remotion. It got pretty close after some back and forth.
And once it got close enough, I had Codex save it as a template so I could reuse it for other videos. Here's the original.
Here's the recreation in Remotion.
And that's what it all comes down to, guys. You find animations you like and ask Claude or Codex to help you recreate them in Remotion. It's just that Claude is doing the editing so you can focus on the visuals.
And if you'd rather skip all of the boring setup, you can just get my engine in my school with the link in the description. I meet every week with my group to train them how to use it and how to sell AI editing services that earn you $2,500 per client.
And I'll talk more about that in a second. I also add new transitions, motion graphics, and other animations regularly, so you can use it to upgrade your engine.
And I recently updated the training so you can create YouTube shorts and reels with it. In fact, I started testing it with my second channel and it's already gotten more subscribers than my main channel. This is freaking cool, and honestly it's really fun to do too.
Which is important if you wanna sell this as a service. So if we haven't met yet, and I guess I should have introduced myself at the beginning, but I get a little nervous on video. I'm Ryan and I've helped my clients earn a collective $2,000,000 by using AI to craft, market, and sell offers.
I first started doing this by selling GPT packages on LinkedIn and just recently started on YouTube. So welcome to my channel.
I sold automations for $2,500 and scaled to 200 k in about ninety days back when they first came out in 2024. So I'm more than confident that you can do the same thing with AI editing.
And then from 2024 to today, AI content automations continue to crush it for me. So I created my school community to teach you how to do the same. Now you can really sell this in two ways.
First, whether you build it from scratch or use mine, you can promote your services by using the automation on your videos. So your offer doubles as your marketing engine which is pretty meta if you think about it. For example, look how I started the video you're watching now.
I said, everything you're seeing right now from the motion graphics, the transitions, and the animations was all done with Claude code and Remotion. This means every video you create is a demo of your services.
So people will see how good your editing is and inquire about your service, but make sure to get targeted with it too. For example, don't just sell generalized AI editing services because it's a race to the bottom with super low hourly rates with people on Fiverr and Upwork.
But get super specific with it like AI video editing for luxury real estate agents who need to sell their listings fast. And this sounds crazy, but it's actually easier to sell at higher rates like 2,500 per month than lower rates like a $100 per hour. It just depends on who you're targeting.
People who buy hourly rates always look for the cheapest possible person, but real business owners care less about price and more about craft. Think about that.
If a luxury real estate agent or even a normal real estate agent has five home listings and they sell even one with your service, they've more than covered the cost. These professionals equate expensive with quality, but low cost with cheap. Plus they will see you as the person who saved them a ton of time from video editing.
You can do this for roofers, plumbers, HVAC, online entrepreneurs, basically any established business owner because business owners need content.
And one last point because you'll love this. Some people will want the one time installation, but others will want you to install it and run it for them.
So you could charge a $2,500 installation fee and then $2,500 per month for ongoing content services.
Or you could add a three month package incentive that reduces the price to $19.97 if they commit to three month increments because this lets you stabilize monthly income.
And if you only got five clients doing that, it's more than a $120 per year. You see how that works? Not freaking bad, dude.
There's a little more to it and it usually raises questions and that's why I created my school community so we can build together every day. Okay, guys. To make this easier, I put the repos for Remotion and WhisperX in the description if you want to build this yourself.
Just tell Claude Code or Codex, you wanna build a content engine project on your computer so AI can edit your videos. Then give it the repo links and tell it to build an end to end video process with Remotion for rendering and Whisper X for audio editing. With some simple back and forth with Claude, you'll have an engine that will let you produce all kinds of content for you and people who need video editing.
If you'd rather skip ahead, I put my school link in the description. And if you want another example of how this works, watch this video next. See you next time.
The Hook

The bait, then the rug-pull.

Ryan opens by pointing at his own screen: the transitions, the zooms, and the layered Vox-style scenes playing right now weren't cut by a human editor, they were directed by Claude Code. He then turns the demo into a pitch, walking through the three-tool stack behind it and the $2,500-per-client service he says it can become.

Frameworks

Named ideas worth stealing.

00:39list

The Three-Tool AI Editing Stack

  1. Claude Code / Codex (director)
  2. Remotion (renders motion graphics in React)
  3. WhisperX (word-level timestamps)

The three pieces of software chained together, with a Python layer telling each one how to talk to the others.

Steal forany automated content-production pipeline that needs code-driven motion graphics synced to speech
02:33model

Three-Layer Vox-Style Animation Prompt

  1. Background
  2. Midground
  3. Foreground

Prompting Claude to design three explicit depth layers, each with its own content and motion, is what produces the layered 3D Vox look.

Steal forany AI-generated motion graphic that needs to feel dimensional instead of flat
07:13concept

Niche-Specific AI Editing Offer

Instead of selling generalized AI video editing at an hourly rate, pick one type of business (e.g. luxury real estate) and price the same service as a flat monthly retainer.

Steal forpositioning any productized service to escape hourly-rate competition
CTA Breakdown

How they asked for the click.

VERBAL ASK
04:52product
you can just get my engine in my school with the link in the description

Dropped right after demonstrating a genuinely useful technique (the reverse-engineered Vox template), so the pitch rides on credibility just earned rather than opening cold.

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

Visual structure at a glance.

open
hookopen00:00
three-tool stack
valuethree-tool stack01:12
vox-style prompt
valuevox-style prompt03:17
skool pitch
ctaskool pitch05:45
pricing close
ctapricing close08:22
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.