A certain clipping channel received 530,000 views in just 7 days.
However, the estimated monthly ad revenue from YouTube was only about $200.
"Is that all?"
If you thought that, please read this article to the end.
Because YouTube ad revenue is not the "main body" of this business.
I'm going to talk about the "backside" of how 530,000 views are converted into $16,000 (2.5 million yen) per month.
Furthermore, I share more in-depth stories and behind-the-scenes info on "AI earning mechanisms" like the one introduced in this article in my open chat "X Operation Strategy Room."
Participation is free, so please take a look first if you'd like.
▼ Open Chat "X Operation Strategy Room"
Let's get started.
"Clipping" Has Become Something Different from the Past
When you hear "YouTube automation," you might imagine faceless videos, cheap synthetic voices, picked-up trivia, and management screens filled with channels no one watches.
That era is coming to an end.
The new way is much simpler.
Choose one creator who is already attracting attention.
Clip the "tastiest" moments of that person's videos.
Post them to all social media platforms.
And connect that "collected attention" to an exit that actually makes money.
This is what a modern "clipping page" is.
The point is here:
A clipping page is not content creation.
It is "distribution."
The Screenshot Everyone is Misreading
Let's take a certain investment-related clipping channel as an example.
This channel is actually not that big. It has about 15,000 subscribers and around 1,100 videos.
It posts clips of a certain creator every day, clearly stating that it is "not the person" and "not an impersonator."
This is important.
This page isn't trying to "be" the creator. It's trying to be the "layer that handles distribution" around that creator.
And there's a part many people overlook in the numbers shown by analysis tools.
530,000 views in 7 days. But the estimated monthly income from YouTube is about $200.
The average video length is 1.6 minutes.
If you think this business is about "YouTube ad revenue," this looks like a waste of time. 500k views a week for only $200 a month? That's not a business; it's unrewarding grunt work.
But there is a "link" in the profile of this page. That is everything.
The clip is not the product.
The clip is "customer acquisition."
The product is what lies beyond (the backend).
Paid communities, online courses, affiliate projects, requests from creators, or companies that pay for the "diffusion" itself.
This is the first rule of clipping pages:
Don't create for the number of views.
Create for the "exit beyond the views."
The Logic Behind the "$16,000 a Month" Calculation
The figure of $16,000 (2.5 million yen) is not magic. It's just a "multiplication of distribution."
If a clipping page can gather 4 million views a month, and the exit pays about $4 per 1,000 views...
When calculated, it becomes about $16,000 per month.
The exact amount doesn't matter. What matters is this "gap."
The exact same attention has a value of only $200 a month in terms of ad revenue, but it becomes worth $16,000 if connected to the right exit.
Same clips. Same views. The only difference is the business model.
There are various exits: tie-ups with creators, content reward programs, affiliate projects, your own sales funnel, or clients who want inquiries rather than ad revenue.
Without this exit, the same attention ends at $200 a month.
This is where most YouTube automation advice misses the mark. They only teach you to chase pennies from the platform and don't teach you the most important thing: "having your own exit."
Money isn't in "posting" short videos. It's in "turning short videos into a flow of money."
Why Claude Opus 4.8?
Now, let's talk about Claude Opus 4.8.
What's amazing about Opus 4.8 isn't that it can "write subtitles." Anyone can do that cheaply now.
What's amazing is that it can act like an "operator (manager)."
According to Anthropic, Opus 4.8 is good at agent-like tasks, has sharp judgment, is strong in long-process tasks, and is designed to move dynamically within Claude Code.
In technical terms, it supports a 1-million-token context, up to 128k token output, situational thinking, and high-speed mode.
Translated into plain English, it means this:
It can remember a lot of information, move according to a large plan, use tools, catch many mistakes, and stay focused on work for a long time.
This is exactly the ability needed for a clipping page.
A human editor creates one clip at a time. But what an operator does is "run the whole machine."
Which creator are we clipping now?
Which videos are still left untouched?
What kind of "hook" is winning?
Which platform needs a post today?
Which clips are already scheduled, and which ones failed?
Which ones should be reposted?
Which accounts are growing?
And which "exit" is this attention connected to?
This isn't an editing problem. It's a "management" problem.
Opus 4.8 is the brain.
Vugola is the clipper.
Postiz is the distribution engine.
When these three are combined, a mountain of a creator's videos turns into a "posting machine."
The "Three Tools" Needed
You need three components.
1st: Claude Opus 4.8 runs the whole thing
You just tell Claude: which creator, which exit, based on which video, what rules, and how many times a week. Claude uses a mechanism called MCP servers to not just "give you a plan" but to "actually do the work."
2nd: Vugola clips long videos
Using the Vugola MCP, Claude can directly operate Vugola. Give it a YouTube video, have it make 9:16 vertical clips, choose the subtitle style, wait for processing, download the finished product, and check usage. Claude does it all. Vugola is an AI clipping tool with a paid plan (from $9/month). When you input a long video, the AI finds "viral-looking moments," formats them vertically, burns in subtitles, and returns them ready to post.
3rd: Postiz handles distribution
Postiz is what schedules and posts the finished clips to each social media platform. Using the Postiz MCP, Claude can operate your social media posting calendar. Instead of leaving finished products in a folder, they are properly put into distribution. Postiz is an open-source distribution tool supporting over 30 social media platforms. It automatically handles the detailed rules of each platform like TikTok, Instagram, YouTube, X, and LinkedIn.
Without Postiz, you just have "inventory." With Postiz, it becomes "distribution."
Setup Procedure in Claude
You only need to set it up once at the beginning. After that, it moves with prompts. Note: Never publish your API keys. Do not paste keys directly into screenshots or social media.
Step 1: Check Node.js
The Vugola MCP requires Node.js version 20 or higher and a paid Vugola account that allows API use. Check the version in the terminal:
1node -v
Step 2: Prepare the Vugola Key
Issue an API key in the Vugola management screen. The key starts with "vug_sk_". Before connecting to Claude, check if the key is valid in the terminal.
Step 3: Connect Vugola to Claude
For Claude Code, the official command is:
1claude mcp add vugola -- npx -y [email protected]
Then, make the key available as an environment variable. (For Claude Desktop, a one-command installer is also available.)
Step 4: Connect Postiz
Go to the developer page from the Postiz settings and copy the API key. In Claude Code, add Postiz via HTTP connection:
1claude mcp add --transport http postiz "https://api.postiz.com/mcp/YOUR_KEY"
Step 5: Confirm both are connected
Finally, restart Claude and check if both servers are visible:
1claude mcp list
Then, try asking Claude:
"List the connected social media accounts." "Check Vugola usage and remaining credits."
If a proper answer comes back, you're ready. Conversely, don't move forward until this is confirmed.
The "Single Instruction" to Move the Machine
Once both are connected, just give Claude "one instruction." That instruction isn't "give me ideas." It's "use the tools and actually do it."
For example:
1You are operating a clipping page for [Creator Name].23The goal is to post short clips and gather attention for [Exit].4Source Video: [YouTube URL]5Use Vugola to create 10 vertical 9:16 clips.67Use a simple subtitle style.8Prioritize moments that move emotions, make people think it's amazing, conflict, surprise, or strong assertions.9Do not choose boring explanatory parts.10Once the clips are ready, download them.1112Next, use Postiz to distribute the best 6 clips across YouTube Shorts, Instagram, TikTok, and X.1314Schedule 2 posts per day for the next 3 days at 9:00 AM and 6:00 PM.1516Write appropriate text for each social media platform.17Finally, show me a calendar with clip names, destinations, and scheduled times.
This is the difference between an "AI tool" and an "AI operator."
A tool "gives" you a clip. An operator "turns the clip into distribution."
What to Let Claude "Judge"
You shouldn't let it clip everything. If you do, you'll end up with a lot of garbage. What you should ask Claude is to "identify the moment."
Viral clips aren't clean explanatory parts. They are moments that make people think it's amazing, moments where emotions move, conflict, surprise, money, lifestyle, pain, or a strong quote that makes people want to share.
This isn't an "education page"; it's a "page that shows aspiration."
And every genre has its own "pattern."
For fitness: transformation, self-management, pain, pride.
For business: money, failure, status, opportunity.
For investment: risk, winning/losing, confidence.
For interviews: confession, conflict, taboo, true feelings.
Don't just have Claude Opus 4.8 look for "parts where someone is explaining something." Have it look for "parts that make people stop scrolling."
30-Day Roadmap
Week 1: Decide on one creator and one exit
Don't choose a "genre." Choose one "creator." Genres are too vague. A creator has a mountain of videos, a voice, fans, an exit, and a continuous supply. Choose someone for whom clipping long videos makes sense. The ideal is having permission, a tie-up, or doing it directly for that creator.
Week 2: Clip and post little by little
Create clips from 3-5 long videos with Vugola. Schedule 2-3 posts a day on suitable platforms with Postiz. Don't flood a new account with 30 posts on the first day. If it looks mechanical, it's because it is. Right now, you are testing "what gets a reaction."
Week 3: Score the winning patterns
Record the hook pattern, theme, which creator's moment it was, platform, views, retention, saves, comments, and clicks. Then, have Claude summarize "what is working." The goal isn't to make the agent creative; it's to create a "loop of improvement so clear it can't be ignored."
Week 4: Scale the winning patterns
If status-type clips win, increase them. If educational ones are failing, drop them without regret. If one platform wins, concentrate there. If the exit starts getting clicks, increase content that answers pre-purchase anxieties. This is how clipping turns from a "hobby" into a "business."
What Kills This Business
Finally, here are the patterns for failure:
No exit: Views without a monetization funnel are just self-satisfaction.
No permission or project: A strategy of just clipping creators without permission is a house of cards.
No volume: One post a day isn't media; it's just a diary.
No sense: Agents only remove labor; they won't make a boring clip interesting.
No distribution system: 10 clips in a folder and 10 clips scheduled on 4 social media platforms are completely different things.
No improvement loop: If you don't record what won, it's just gambling.
The Most Important Lesson
Claude Opus 4.8 didn't just "make YouTube automation easier." It changed "what YouTube automation means."
The old game was: make a video, upload it, and pray for ad revenue. The new game is: find attention, clip it, distribute it, and flow it to an exit.
Vugola handles clipping. Postiz handles posting. Claude runs the loop. Still, the human operator needs "judgment." Which creator to choose? Which exit to choose? What kind of attention is worth chasing? That's for you to decide. But once decided, the machine doesn't need your hands all day. Therein lies the opportunity.
Not "faceless YouTube automation," but an "agent-run clipping page." A page with 15,000 subscribers might look small from the outside even if it gets 530,000 views a week. But to those who understand the backside, the "assets" beneath it are visible: acquisition assets, distribution layers, and a "media machine waiting for an operator."
This was a translation and my own commentary on an article that was viral overseas. This story looks like it's about clipping, but the essence is much deeper. I think the most important sentence was this:
"Don't create for the number of views. Create for the 'exit beyond the views.'"
This isn't just about clipping. It's the same for X operations, blogs, and affiliates. Too many people only think about gathering attention and don't think about "where to flow the gathered attention." The same attention can be worth $200 or $16,000 depending on the exit.
And one more thing: whether you use AI as a "tool" or an "operator." The difference in this mindset will become a massive gap from now on. If you just have it write subtitles, it's a tool. If you have it run the whole machine, it's an operator. How far you can expand your thinking when new tools come out—that, in the end, is everything.
Finally: Sharing the "Behind the Scenes" of Earning with AI
I share more in-depth stories and realistic information that's hard to put out publicly about "new ways to earn using AI" in the open chat "X Operation Strategy Room."
・Latest examples of AI x Monetization currently viral overseas
・The backside of posts that actually got a reaction
・Realistic numbers and verification data not published on X
Furthermore, I'm currently developing a "tool" to accelerate X operations, and once completed, I'll provide advance guidance in this open chat. Participation is free. Only those who want to "move one step ahead with AI" should take a look.
▼ Open Chat "X Operation Strategy Room"
Let's ride the new wave faster than anyone else together. I look forward to your participation.
Ryuji Yamamoto (@ryuji_affiliate
This article is a translation, reconstruction, and commentary based on overseas articles and public information of each tool. Amounts are estimates based on 150 yen to the dollar, and tool specifications/fees fluctuate, so please check the latest information on each official website before use.
