11:09Karolis · Tutorial
The 2 Claude Skills That Run My Entire AI UGC Video Pipeline
How a UGC Decoder skill and TopView Canvas replace a multi-day production workflow with a single reusable session.
June 10thThe argument in one line.
A small set of pre-built Claude skills collapses motion graphics, carousel design, video editing, and viral trend research into a single AI interface — no long prompts, no extra subscriptions, no designer required.
Read if. Skip if.
READ IF YOU ARE…
- You post short-form content on Instagram, TikTok, or YouTube Shorts and want to systematize production without hiring a designer or editor.
- You already use Claude but write one-off prompts each session and want reusable skills that carry context.
- You spend meaningful time cutting silences, building carousels, or hunting for content ideas from scratch.
- You want studio-quality motion graphics without outsourcing or paying for a dedicated animation tool.
SKIP IF…
- You are not actively creating content — this is a production workflow toolkit, not a strategy video.
- You need frame-accurate, multi-track video editing; the silence-removal skill is a single-prompt blunt instrument, not a full NLE.
The full version, fast.
Claude's built-in knowledge can't tell you what's actually trending right now — it hallucinates. This video introduces a six-skill toolkit that patches that gap and automates the rest of the content stack: connect Virlo as a live data source for real post performance, run a validated hook through the carousel skill to get a full slide outline, use the GPT Image 2 prompt skill to generate reference-matched visuals for free, drop raw footage into Claude Desktop to strip silences in minutes, and activate the static visual or motion graphics skill to produce polished on-screen graphics from a single script line. All six skills are free to download; Virlo is pay-per-use from $10.
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 →Where the time goes.
00:00 – 00:27
01 · Introduction
Hook claim and rapid montage of motion graphics, carousel, and video editing outputs as proof of concept.
00:27 – 00:46
02 · Skill Setup
Download zip, open Claude, Customize → Skills → upload each file. 30 seconds total.
00:46 – 01:40
03 · Motion Graphics Animator
Activate skill, paste one script line, receive complete Opus 4.8 prompt. Paste into Opus, get HTML animation. Screen-record or convert to MP4.
01:40 – 03:42
04 · Virlo: Live Trend Data
Claude hallucinates trending content. Connect Virlo as a custom connector for real post data — view counts, hooks, patterns across 4M+ short-form posts.
03:42 – 05:07
05 · Build Instagram Carousels
Virlo-validated hook → carousel skill → full slide outline. GPT Image 2 prompt skill → reference photos → production-ready image prompts → generate free in browser.
05:07 – 06:16
06 · Video Editing (Silence Removal)
Claude Desktop required. Drop MP4, prompt 'remove all silent parts longer than 0.5 seconds.' ~10 min. Must be MP4 or audio sync breaks.
06:16 – 08:00
07 · Create Static Visuals (+ Motion Upgrade)
Static visual skill turns one script line into a polished HTML graphic. Add one line for vertical version. Chain into motion graphics skill for full animation.
Lines worth screenshotting.
- Claude hallucinates when asked what is trending in your niche — it has no live access to social media data.
- Connecting Virlo as a custom connector in Claude takes two minutes and gives real view counts and hook analysis for 4 million tracked short-form posts.
- A single Claude skill can turn one script line into a production-ready Opus 4.8 motion graphics prompt — no animation knowledge required.
- GPT Image 2 is free in the browser; generate 20 images at once and expect roughly 5 usable ones per batch.
- Claude Desktop's video editor skill removes silences from raw footage with a single prompt, eliminating what used to be a two-hour manual edit.
- Generating 20 AI images at once costs the same as generating one — batch generation is the correct workflow, not one-at-a-time.
- A static visual made from one skill can be chained directly into the motion graphics skill for a full animation with no additional prompting.
- Claude skills installed once remember their context for every future session — zero prompt re-writing.
- The silence-removal skill requires MP4 format; any other format breaks audio sync.
- Claude's carousel skill builds the full slide outline first, then the GPT Image 2 prompt skill handles the visuals — two separate skills, one integrated workflow.
Six skills that replace a full production team.
WHAT TO LEARN
Pre-built Claude skills eliminate the per-session prompt tax and close the biggest gap in AI content tools — lack of live data.
- Claude skills are reusable JSON files that give Claude persistent context, so you type a name instead of re-writing a long prompt every session.
- Claude cannot tell you what is currently trending on social media — it hallucinates. Connecting a live data source like Virlo via a custom connector is the only reliable fix.
- Generating images in batches of 20 is the correct workflow for AI image tools — the yield rate is roughly 5 usable images per 20 generated, and batch cost is the same as single.
- Silence removal in Claude Desktop works well for solo footage in MP4 format; any other container breaks audio sync — format matters more than the prompt.
- A static visual generated from one skill can be passed directly to the motion graphics skill to produce a full animation, treating outputs as modular inputs across skills.
- The full 6-skill content stack — trend research, carousel design, image generation, video editing, and motion graphics — is available free; the only paid component is Virlo's pay-per-use API balance.
Terms worth knowing.
- Claude skill
- A reusable JSON configuration file uploaded to Claude that gives it persistent context, instructions, and behavior for a specific task — so you activate by name instead of re-writing a long prompt each session.
- Virlo
- A short-form content intelligence platform that tracks over 4 million posts across TikTok, Instagram, and YouTube Shorts in real time. Integrated into Claude as a custom connector to return real performance data instead of hallucinated trends.
- Custom connector (Claude)
- A live data source linked to Claude via Settings → Connectors. It lets Claude query an external API in real time rather than relying solely on its training data.
- Claude Desktop (co-work)
- The downloadable desktop app version of Claude that supports local file uploads, including video files. Required for the video editor skill because the browser version cannot accept video.
- GPT Image 2
- OpenAI's image generation model, free to use in the browser at chat.openai.com. Used in this workflow after the Claude skill generates a detailed image prompt from reference photos.
Things they pointed at.
Lines you could clip.
02:10
“If you go to Claude and ask what is trending in your niche, it just makes stuff up.”
04:00
“Not all of them are going to come out usable. Some are going to look slightly off, but if you generate 20 at once, at least five are going to be exactly what you need.”
01:40
“Animations like this used to cost me a fortune to outsource, and this skill changes that completely.”
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.
00:00So I spent weeks building Claude skills that completely change how you make content. These skills help you make motion graphics like this, edit your videos, design Instagram carousels like this, and a lot more. You no longer need long complex prompts.
00:16Just install them once and you're ready to go. And I'm giving away all of them for completely free so you can start using them right away. All I ask in return is for you to drop a like on this video.
00:26Now let me show you how they work. Before anything else, you need to install these skills. Just download the zip file from the link in the description and extract it.
00:34Open Cloth. Go to customize, then skills. Hit the plus button and create skill.
00:39Click upload a skill and upload each file one by one. Thirty seconds and you are done. Now let's use them.
00:46Animations like this used to cost me a fortune to outsource, and this skill changes that completely. Once it is installed, Cloth already knows how to build cinematic animations like this. So first, just activate the skill by typing this, and then all you need to do is take a line from your video script, drop it in with this prompt, and hit send.
01:07And we get a complete ready to use prompt. Now I select the Opus 4.8 model, drop this exact prompt that we just got, and wait couple of minutes.
01:16And we get this. Since it comes out as an HTML file, you have two options. You can screen record it directly in your browser or use a simple HTML to m p four converter.
01:28Now if you need animation for a vertical short form content, just use this prompt instead. Same process. Just drop your script in, get the prompt back, select Opus 4.8, and a couple of minutes later, you get this.
01:41Now this next one is going to find you viral content ideas in minutes. I use this every single day for my own Instagram content, and it is a big reason some of my posts hit 600,000 likes like this one right here. And I also have other posts that are performing just as well.
01:57And to find ideas for all of these posts, I always use this skill. This one works a little differently from the skills that we just installed. Instead of a file, we connect it directly to cloth as a live data source.
02:09Because if you go to Cloth and ask what is trending in your niche, it just makes stuff up. It has no real data on social media landscape, but you can fix that by connecting Cloth to Verlo. Verlo tracks over 4,000,000 pieces of short form content across TikTok, Instagram, and YouTube shorts in real time, and it takes two minutes to connect it with Cloth.
02:29Go to the first link in the description and log in to Verlo. Head to API keys, generate a new key, then open Cloth. Go to settings, connectors, add custom connector, paste the Verlo URL, hit connect, paste your key, and authorize.
02:42That is genuinely it. Let's say I want to find what is actually performing right now in the self improvement niche. So I would just type something like this and hit send and look at what comes back.
02:54These are not some made up ideas. These are real posts with real view counts and real breakdowns of why each one work. It shows you the hook, the angle, and why it performed so well.
03:05This one, for example, has 7,900,000 views on an account for just 27,000 followers, which is a clear outlier.
03:11And down here, Verlo breaks down the patterns across all of the top performers. And you can find content ideas for nearly every major niche, so you will generally never run out of ideas. I can just switch from self improvement to travel niche and get the same real time data instantly.
03:26And instead of spending hours scrolling trying to figure out what content works in your niche, you just connect Verlo to cloth and get real data in seconds. You can top up your Verlo balance for as low as $10 and start using it right away. The Verlo link and all the skills from this video are in the description down below.
03:42So now we have a proven content idea with real data behind it. Let me show you how to turn it into a full carousel using the next skill. First, we turn the Verlo idea into a full slide outline.
03:54So let's activate the carousel skill and then drop this prompt. The hook we're gonna use is basically already done because Verlo told us it works. I just changed it a little bit with cloth and went back and forth a couple of times on the rest of the slides until I had the final result.
04:08This is where the carousel skill came in and handled all of that for me. Now the question you're probably asking right now is how did I make these photos look that real? So this is where the GPT image two prompt skill comes in.
04:21Find 10 to 20 reference photos that match the vibe and aesthetic you want to go for. Activate the GPT image two prompt scale and just drop all your reference images into Claude with this prompt. Then I take those prompts straight into GPT image two and drop them in one by one.
04:38And look, I'm gonna be real with you. Not all of them are going to come out usable. Okay?
04:42Some are going to look slightly off, but if you generate 20 at once, at least five are going to be exactly what you need. Just keep generating until you get the ones that work. GPT image two is completely free to use in your browser, so no extra subscription is needed.
04:57Once you have your visuals, take the outline cloth built, drop your hook and value text into each slide, and you have a full carousel ready to post. Okay. So this next skill is going to save you a lot of time.
05:10Look. I don't know a single creator who enjoys this part. You finish recording, you're feeling good about it, and then you open the timeline and it's full of silence parts that you have to cut out.
05:22And you spend the next two hours clicking through removing every single one. So let's activate the video editor skill by typing this. And what the skill does is you drop a raw footage into cloth co work and you say, edit this video and remove all silent parts longer than zero point five seconds.
05:40And that's it. You need co work for this one specifically because you can't upload actual video files into Cloth in the browser.
05:47So just download Cloth co work, open it, drop the video file that you want to edit, and send that prompt. Just make sure the file you're using is in m p four format. Anything else and the audio syncing is going to be off.
05:59And it just starts working on it. Ten minutes later, here's what we get. Plan your day the night before.
06:06Go to sleep and wake up at the same time. Plan your day the night before. Go to sleep and wake up at the same time.
06:13Clean file with all of the silences removed. You know these clean visuals you see in YouTube videos where somebody is explaining some sort of concept and there is just a polished graphic on screen that perfectly illustrates what they are saying. Those visuals used to take me twenty minutes to make every single time, and I'm talking just one visual.
06:33Now I just activate the static visual skill, take the line from my script where I need a visual, drop it into cloth and type create a clean static visual for this part of my script. Then you just paste your script in these brackets.
06:49And in couple of minutes, we get this. And if you need the same visual for vertical content, just add one line for this prompt, which explains that you want a different format this time. And now we have both formats ready to go.
07:01You can do this for every single part of your video that needs a visual, and you end up with a full library of graphics ready to use before you even sit down to edit. Another cool thing you can do is literally turn that static visual you just generated into a full blown motion graphic animation. Just type, I want to turn the static visual into a motion graphic animation, and Claude will generate a prompt you can take straight into Opus 4.8.
07:26Open a new chat, select the Opus 4.8 model, drop in the prompt, and in a couple of minutes, you get a full blown animation for your static visual. And if you want, you can do the same thing with all other static visuals that you generate. Honestly, it's crazy how fast AI is moving right now, and we are just getting started.
07:45If this saved you any time, make sure to subscribe because I'm building more of these skills and sharing them with you for free. And if you want to see more of what Cloth can do for your content, make sure to watch this video right here. Thank you so much for watching, and I'll see you in an
The Hook
The bait, then the rug-pull.
Weeks of prompt engineering, packaged into six files you install once and never rewrite. This breakdown documents the exact workflow — from motion graphics to silence removal — that a single creator used to replace an entire production stack with Claude.
Named ideas worth stealing.
00:27list
The 6-Skill Content Stack
- Motion Graphics Animator
- Virlo Trend Research (live connector)
- Carousel Builder
- GPT Image 2 Prompt Generator
- Video Editor (silence removal)
- Static Visual Creator
Six Claude skills covering the full short-form production workflow — install once, activate by name, no long prompts.
Steal forAny content system walkthrough or AI toolkit offer
02:37model
Virlo Connector Setup
- Log in to Virlo → API Keys → Generate new key
- Open Claude → Settings → Connectors → Add Custom Connector
- Paste Virlo URL → Connect → Paste key → Authorize
2-minute integration that replaces Claude's hallucinated trend data with real post performance from 4M+ tracked videos.
Steal forAny tutorial showing how to connect live data sources to Claude
How they asked for the click.
VERBAL ASK
07:35subscribe
“If this saved you any time, make sure to subscribe because I'm building more of these skills and sharing them with you for free.”
Soft, value-anchored ask. Frames future content as the reward for subscribing rather than just asking for a follow.
MENTIONED ON CAMERA
FROM THE DESCRIPTION
AFFILIATECommission earned if you click.
OTHER LINKSAlso linked in the description.
Visual structure at a glance.
