YT YouTube2Text

YouTube2Text → Transcripts → Как Я с Claude Сделал $32,740/месяц на ИИ YouTube Канале (пошагово)

Как Я с Claude Сделал $32,740/месяц на ИИ YouTube Канале (пошагово) — Transcript

by Shukur Nabiev | YouTube · 4,321 words · 624 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:00Now I will show you how much my YouTube
0:02channels actually earn. Here is the
0:03first channel. We launched it last year
0:05, and in the last 28 days, it brought in
0:09$ 9,892. Refreshing the page, that comes
0:11to 769,000 rubles. The second channel
0:13works using the exact same system and
0:15concept. Here, in the last 28 days, we
0:18have $ 8,235 from monetization. In the
0:20other currency, that’s 663,000 rubles.
0:23We started this channel about six months
0:25ago, and it’s already making $ 7,648 a
0:27month. In rubles, that is 615,000.
0:30And the last channel brought in $ 6,965
0:33over the past month. Changing the
0:35currency, that’s 560,000 rubles. If you
0:38add it all up, these four channels
0:40brought in $ 32,740 in one month. The
0:42craziest part is that producing this
0:44content takes me 2-3 hours a day. Since
0:46this works so well, I’ve already
0:47launched another channel with the team
0:49using the same system.
0:50Today I will show it to you entirely,
0:52step by step, with no fluff, so any
0:54beginner can do the same, even faster
0:56than I did.
0:57But let me say upfront: I am not selling
0:58any private courses. There is no "get
1:00rich quick" button here. I am not saying
1:01you will succeed overnight. This is not
1:03a scheme to get rich with a snap of your
1:04fingers. And the result will depend on
1:06how much effort you put in. But for all
1:08of this to work for you, you need to
1:10avoid this kind of result. Zero views.
1:12Some people call it a shadowban. I often
1:14hear pseudo-experts on YouTube saying
1:16that zero views happen because the video
1:19was made by AI. And it's not always
1:20interesting. Do they even understand
1:22what they're talking about? Because that
1:23is complete nonsense. Here is a clear
1:25example of one of our long-standing
1:26channels. You can find it on YouTube
1:27yourself.
1:28The voice here is AI-made, the images
1:30are AI-made, the script is AI-made,
1:31everything is done through AI.
1:33And as you can see, we have been making
1:35this content for several years. It was
1:36monetized back then and it is monetized
1:38now. Here are the videos, here are the
1:40views—it is exactly as it is. And this
1:41situation applies not only to us. Here
1:43is a very young channel, Ark Survival
1:45Stories. According to VidIQ data, it was
1:47created in March.
1:48The videos are marked as AI-made, and
1:50yet in mid-October, it had 74
1:52subscribers and about 15,000 total
1:54channel views.
1:55But in the last 2 weeks, it grew to
1:571,500 subscribers and more than 300,000
2:00views. If you run those views through
2:02even the VidIQ calculator, with an
2:04American audience, it comes out to about
2:06$ 1,000 to $ 3,000 per month. And since
2:08we make such channels ourselves, the pay
2:10per 1,000 views here is about $ 7.
2:13Accordingly, the creator earned at least
2:15$ 2,100.
2:16So, when the next pseudo-expert says
2:17that AI doesn't monetize, just don't
2:19listen and follow the practitioners.
2:21People are entering this niche, and many
2:23others are already earning money right
2:24now. Alright, let's figure out how not
2:26to end up with zero. First thing to
2:28check is whether your account is even
2:29verified. Open your browser and enter
2:31YouTube dot com slash verify. See, mine
2:33says that the account is verified. If
2:34yours doesn't, here is what we do. Go
2:36into YouTube Studio and open the
2:37settings. Next is the video uploads
2:39section and this tab right here. Set
2:40this to "unlisted" access. I’ll explain
2:42a bit later why it should be unlisted
2:44and not public. For now, just keep that
2:46in mind. Now go to the channel tab and
2:47open feature eligibility. There are
2:49three levels here. The first one is
2:50always verified for everyone right away
2:52. But the other two you need to handle
2:53yourself. Click the arrow, and a button
2:55appears asking you to verify your phone
2:57number. Click it, enter your number or a
2:59temporary one from any website. A code
3:01arrives; enter the code and click verify
3:03. That's it, the second level is ready.
3:04And only the last one remains. Mine is
3:06verified through channel history. This
3:08means I’ve built up data on the channel
3:10by regularly uploading decent videos.
3:12Not trash, but proper content. And you
3:14will learn exactly how to make that
3:16today. If you don't have history yet,
3:17you can leave this for later. It will
3:19automatically verify itself over time.
3:21Or you can verify it immediately with
3:23documents or just record a short video
3:25explaining your channel and upload it.
3:26Once you’ve done everything, come back
3:28here, and all three levels will be
3:30verified. But that’s not all. Open the
3:32advanced settings tab and select the
3:33middle option. I’ll explain now why this
3:35is important. The first option sounds
3:37like you’re saying your content is safe
3:39and clean, with no swearing or adult
3:40themes. But on YouTube, the "made for
3:42kids" label doesn't mean it’s
3:43family-friendly at all. It means the
3:45video is made specifically and only for
3:47children under 13. Meaning cartoons,
3:49kids 'songs, or, for example, toy
3:51reviews. If you check this box for
3:53regular content—the kind we'll be making
3:55today—YouTube effectively puts your
3:57channel into lockdown mode. Let me
3:59explain now why this kills your channel
4:00. Look, as soon as you check that box,
4:02YouTube completely disables comments.
4:04This means you can’t interact with your
4:06viewers, they can’t ask questions, and
4:07you get no feedback. The channel turns
4:09into a complete ghost town. And here is
4:11the most dangerous part. The algorithm
4:12gets completely confused and starts
4:14showing your video to seven-year-olds in
4:16the YouTube Kids app. The kids see it’s
4:18not a cartoon and leave immediately. For
4:19the algorithm, that’s a signal that the
4:21video is bad, and it stops showing it
4:22altogether. And do you know what’s even
4:24more unpleasant? YouTube disables
4:25everything that brings viewers back. The
4:27notification bell stops working, which
4:29means your subscribers—the people who
4:30like your content—simply won't get
4:32notified about a new video. Videos
4:34cannot be saved to watch later. The
4:36mini-player is disabled. If a person
4:37minimizes the video to scroll through
4:38YouTube, the video will simply stop. And
4:40, of course, you won't be able to add
4:42end screens either, which help gain
4:43watch time. Therefore, always choose the
4:45second option, stating that your content
4:46is not made for kids. Scroll a little
4:48further down, and you will see automatic
4:49dubbing here. There is a button next to
4:50it that prompts you to turn it on. Just
4:52turn it on. Then open the basic info
4:54section. You need to paste keywords here
4:56that perform best in your niche. If you
4:58open YouTube in Incognito mode and type
4:59, for example, "History" into the search
5:01bar, a bunch of suggestions will pop up
5:02immediately. These are queries that are
5:04constantly searched for, but typing them
5:06in one by one is pointless. Besides, you
5:07won't know if they are worth pursuing
5:09because you can't see the search volume
5:11or the competition. For this, you will
5:12need vidIQ. Honestly, if you are serious
5:15about creating content, this is the
5:16first and simplest tool you should
5:18install. It has a free plan. Log in
5:20using the same email your channel is
5:21registered with, and you will land on
5:23this page. Click on "Research" and type
5:25"History" into the search bar. And here
5:27, pay attention, we are creating an
5:28educational channel. Therefore, the same
5:30idea can be applied to food, health and
5:32fitness, gaming, and even technology.
5:34These niches also have a very good
5:36audience. We, for example, do history,
5:38but my history isn't about Alexander the
5:40Great and things like that. For instance
5:42, on one of my channels, I tell the
5:43history of games, food, true crime,
5:45survival, and much more. That is the
5:47beauty of historical niches. You can
5:49take almost any topic and tell its story
5:51. That’s why I can recommend it to
5:52beginners. The task here is simple. We
5:54need related keywords with high search
5:56volume and low or medium competition.
5:59But if the competition is high, the
6:00query should have at least 500,000
6:02searches. Look at these words. Almost
6:04all of them have low or medium
6:06competition, and for those where it is
6:07high, the volume is good. Select at
6:09least five, click here to copy them, and
6:11paste them into this field. Click save.
6:14Next, open the branding tab and upload
6:16your logo. You don't need to
6:17overcomplicate the banner. Go to Canva.
6:19The free version will be more than
6:20enough. Sign up, switch to templates,
6:22and search for a YouTube banner. There
6:24will be a lot of results. But if you
6:26want your banner to display properly on
6:28all devices, do this. Pick a template
6:30with these kinds of frames. And look,
6:32there are paid templates here. If you
6:33are on the free version, don't use those
6:35, otherwise you won't be able to
6:36download it later. I'll take this one
6:37and click "Edit template." As I said, no
6:39need to overcomplicate things. One
6:41inscription is enough for a banner.
6:42Subscribe. But it's important that the
6:44text fits into this small area; it’s the
6:46only part visible on all devices,
6:48including phones. Then you can remove
6:50the background and set any simple color
6:52, like here, or any background you like
6:54. The main thing, again, is not to use
6:55paid elements since you are on a free
6:57plan. Click "Share," then "Download,"
6:59then "Download" again to save the image
7:01, and just upload it here. Now for the
7:03channel description. It shouldn't be
7:04long. The main thing is that it contains
7:06plenty of keywords. I have a separate
7:08master prompt for this. For your
7:09convenience, I will post it, along with
7:11all the prompts from this video, in my
7:13Telegram channel comments. I will also
7:15leave the link in the description. You
7:16can go there, take a look, just copy it
7:18, and follow along with me. We copy the
7:20prompt for the channel description and
7:21go to Claude or ChatGPT. They are free
7:23as well. We paste the prompt into the
7:25chat and insert our keywords there too.
7:27Click "Send," and that's it, the whole
7:28channel description is ready. Copy it
7:30here, paste it into the description
7:32field, and click "Publish." That's it,
7:33the channel is ready to go. And now the
7:35dangerous part begins. You can set up
7:37your channel perfectly, do everything I
7:39showed, and still get absolutely zero
7:41views. The craziest part is that you can
7:43spend hours on a video and not even
7:45realize the mistake was made before you
7:46even started. And we don't know what
7:48we're waiting for. There is one thing
7:49that determines in advance whether a
7:51video will get 1,000 views or if no one
7:52will notice it at all. And most
7:54beginners fail at this even before they
7:56start. It is the video idea itself.
7:58Honestly, this is the ultimate YouTube
7:59cheat code, and it's much easier than
8:01trying to guess what people are
8:02interested in.
8:03So, let me show you how I find ideas
8:05that already have the potential to
8:06literally blow up and get many views
8:08before I spend even a minute on them.
8:10We go to vidIQ again, open it, and enter
8:13our keyword. My channel is about history
8:15, true crime, and survival. I cover many
8:17topics, but these work best so far, so I
8:19enter, for example, "how to survive." I
8:22could just write "how to survive a lion
8:24attack" or something similar, but I
8:25don't want to guess. I want to see which
8:28survival topics people are actually into
8:30. Open vidIQ. There are tons of videos
8:32here, but how do you choose a topic that
8:34can really take off? I have two rules
8:36that I follow every time. It is
8:37precisely because of them, among other
8:39things, that my channels grow faster,
8:41because the topics I find always bring
8:42in good views. Rule number one: I look
8:44for small channels whose videos are
8:45getting a lot of views. Rule number two
8:47: I check whether the topic continues to
8:50gain views over time. I'll show you now
8:51. Click on filters. Since we need small
8:54channels that have blown up, I set the
8:55maximum number of subscribers to 5,000.
8:57To have more topics. I click here and
8:59select all countries. I set the Outlier
9:01score to 50 or higher. And since I need
9:03topics that continue to gain views, I
9:05also set the second metric to 50 or
9:06higher. For views, I don't take anything
9:09below 50,000.
9:10Remember that we need a viral topic.
9:12Click apply. Look at this result. A
9:14video about how to survive a zombie
9:16epidemic from day one to day thirty. The
9:18channel has only 1,000 subscribers, but
9:20the video gained over 130,000,000 views
9:22in a week. Outlier score 70, meaning the
9:24video is 70 times stronger than the
9:26channel's usual videos. This means the
9:28viewers are genuinely interested in the
9:29topic. This means if we take this topic
9:31, we can get a similar result or even
9:33better, because the video itself is
9:35so-so. Look at the engagement rate, it's
9:36less than 1%.
9:37If you make a decent video on this topic
9:39, the chances are very good, and you
9:41will be surprised at the video we can
9:42put together in just a few clicks. But
9:44there is one thing you need to know. The
9:46topic you take from here determines
9:47which countries your views will come
9:49from. For example, here is a video about
9:51life in India and the biggest slums. The
9:53channel has 2,500 subscribers, and the
9:55video has over 300,000 views. And guess
9:57what, one of the keywords is India. This
10:00means if we make a video on this topic,
10:01we are targeting viewers from India. And
10:03not just from India. There are a bunch
10:05of nearby countries that are also
10:06interested in how these people live.
10:07Bangladesh, Nepal, Pakistan, and so on.
10:09And this has a good side and a bad side
10:11. The good thing is that the channel can
10:13take off very quickly because India has
10:15one of the largest audiences on YouTube
10:17. But the bad thing is that you will
10:18earn very little from advertising.
10:20There's no need to even explain that;
10:21you already know everything. That's why
10:22you should try to avoid getting involved
10:24with low-rate countries. So let's use
10:26our heads and some logic. Instead of
10:28life in Indian slums, let's look at the
10:30USA, Britain, or, say, Spain. I mean, do
10:32you understand what I'm talking about?
10:33You don't just need to mindlessly copy.
10:35There is one more thing that almost no
10:36one talks about. If you want to grow
10:38specifically with an American audience,
10:39we will get to that, don't worry. Look
10:41at what I do when I find a topic. I copy
10:43it from here, open Google Docs, and
10:45create a new document. Next, I go to
10:47Insert, then Table, and choose how many
10:49columns and rows I need. By the way, the
10:51template for this table is in my
10:53Telegram channel, don't worry. The first
10:55column. I write down the channel's
10:56subscribers. Next is the topic, then
10:58views, views per hour, and the outlier
11:00score. And here is what I do. I spend a
11:02whole day searching for topics and then
11:04fill in everything I found into this
11:06table. Let's look at a real example. The
11:08channel has 1,000 subscribers. I write
11:10that down. I copy the topic and paste it
11:12here. Views per hour are 167, with over
11:14130,000 total. The outlier score is 70.
11:17Returning to VidIQ. Here is another good
11:19topic, also about how to survive in a
11:21zombie apocalypse. The channel has 1,300
11:24subscribers, and the video has over
11:26113,000 views in less than a month. And
11:28look at the views per hour, about 30.
11:30The video is already dying out, and
11:32engagement is low. This means we can
11:33take advantage of this and make a better
11:35video. We add this topic to the table as
11:37well. You can collect as many of these
11:38topics as you want, so you won't have to
11:40struggle with finding ideas anymore.
11:42When you're ready to make a video, just
11:43open the table, pick a topic, and get
11:45started right away. So, we have an idea
11:47, but before making the video, you need
11:49to understand one thing. Just because a
11:51topic got millions of views doesn't mean
11:53your video will too. You can take the
11:55exact same topic that hit a million,
11:57make your own version, and struggle to
11:59get even a hundred. And this is where it
12:01gets interesting. Two creators talk
12:02about the same thing, but one keeps
12:04people until the very end, while the
12:06other loses almost everyone in the first
12:0730 seconds. What is the difference? It
12:09all starts with how the script is
12:11written. There is one specific mistake
12:13that must be avoided. If you make it,
12:15viewers will leave before the most
12:16interesting part. Look, those two videos
12:19we found could have reached a million.
12:22People like the topic, but there is no
12:24emotion in the videos themselves, which
12:25is why engagement is low. Let me show
12:27you how we write a script. I have a
12:29master prompt, and I'm giving it away
12:31for free. It's also in the TG channel.
12:32It writes a script as if you are
12:34actually telling a story to a friend. By
12:35the way, do you know why a bunch of AI
12:37channels got demonetized? It’s actually
12:40a bit funny, but we’ll get to that soon
12:41. I copy the prompt and open Claude.
12:44Note that I am using the free plan. I'm
12:46showing this on purpose so you start
12:47doing something instead of just watching
12:49. I paste the prompt into our chat. I
12:51scroll up. And here you need to enter
12:53the topic. I paste one of the topics
12:54from our table. Next is the duration. I
12:56recommend 15 minutes or more here.
12:58That's about 2,300 words, but I'll just
13:01write 15 minutes. Let's send it and take
13:03a look at the script. These aren't those
13:05cookie-cutter texts that don't hook
13:06anyone. And now, the main thing.
13:08YouTube might not enable monetization on
13:10your channel because of an AI voice or
13:12the AI tools used to make the video, but
13:14because of the content you share.
13:16If the algorithm realizes you are
13:18spreading completely false, nonsensical
13:19information, it will disable your
13:21monetization. And these are neural
13:22networks. We know they make mistakes
13:24constantly. So, there are two options:
13:26either check everything manually or ask
13:28the AI to check itself. But simply
13:30writing "Please fact-check this" won't
13:32work. You need a proper prompt. I am
13:34providing that as well so you don't burn
13:35out on manual verification. In exchange
13:37, I ask that you support my channel with
13:38a subscription. I copy the verification
13:40prompt, feed it into Claude, and wait.
13:42Look, it checked the entire text and
13:44found all these claims. Only a few of
13:46them turned out to be true, and all
13:47these others are misleading. So tell me
13:49, is YouTube to blame if it disables
13:51your monetization for videos with such
13:53terrible scripts? Of course not. I don't
13:55blame you either. You simply didn't know
13:57about this. Here is the corrected
13:58version. Now our script is ready. Next,
14:00let's turn it into a video that will go
14:02viral. Pay close attention here, because
14:04there is one mistake I constantly see
14:06beginners make. They spend so much time
14:07on one video that they eventually burn
14:09out and stop posting. And the most
14:11frustrating part is that most of this
14:12work didn't even need to be done. Look,
14:14there are two ways to make a video. The
14:17first one I call manual. You generate
14:18the images yourself, do the voiceover
14:20yourself, and assemble everything in
14:21CapCut into a finished video. I
14:22recommend doing the voiceover in this
14:24service. You can generate unlimited
14:26voiceovers there at the lowest prices on
14:27the market. Just choose a
14:29natural-sounding voice, paste your text
14:30, and in a couple of minutes, you have a
14:32ready track. I will also share the link
14:33in the description. There is nothing
14:35wrong with the manual method, but
14:36imagine having to go through this
14:38process every single time you want to
14:40post a video. It is quite a lot of work
14:42, which is why we started looking for
14:43another method that would take care of
14:45almost all the routine so you don't
14:46spend hours assembling everything by
14:48hand. For that, there is this service.
14:50It’s Forge Vision; this is the second,
14:52automated method.
14:53Instead of generating everything
14:55separately, you turn your script into a
14:56finished video with one click.
14:58Let me show you how it works. Follow the
14:59link in the description and you'll end
15:01up here. Take our script from Claude and
15:03drop it into Forge Vision. Next, choose
15:05what kind of video it should be: with
15:06photos, with stock footage, or with
15:08animation. Essentially, you decide the
15:10format, and the service assembles it
15:11based on the meaning of the script. You
15:13can make it any length you need—20
15:15minutes, 30 minutes, or even several
15:17hours long. The service automatically
15:18breaks the text into scenes and selects
15:20visuals for each one. We click start. It
15:22takes very little time to make such an
15:24episode, and you don’t have to sit in
15:25front of the screen building every scene
15:27by hand. The service generates the
15:29images, handles the voiceover, edits,
15:31and puts it all together into one video
15:33. You could even queue up 10 scripts
15:35overnight, wake up, and have 10 videos
15:36ready to go.
15:38And most importantly, it’s dozens of
15:39times cheaper than the alternatives.
15:40Once the video is ready, all you have to
15:42do is download it and upload it to your
15:44channel.
15:44>> In August 1914, the ship Endurance
15:47departed from Plymouth.
15:49On board were 28 men and Captain Ernest
15:51Shackleton.
15:52The goal: to be the first to cross
15:54Antarctica on foot. Five months later,
15:57the Weddell Sea was locked in ice. The
15:59ship got stuck and drifted with the ice
16:01for 10 months, until the ice finally
16:03crushed it like an eggshell.
16:05The men were left on an ice floe in the
16:07middle of the ocean. Without a ship,
16:08without communication, in temperatures
16:10of minus 30.
16:10>> I think that’s just amazing. And it was
16:12all put together in just a few clicks.
16:14After we made videos on Forge Vision, we
16:16started a channel where all the videos
16:17are made exclusively through it. And
16:19here is what it has already achieved.
16:20More than 2,300 subscribers and over
16:238,300 hours of watch time. You need
16:251,000 subscribers and 4,000 hours for
16:27monetization. So we passed that mark
16:29with ease and have already earned this
16:31much from it. Now, the thumbnails. We
16:32have our own method for this. We open a
16:34competitor's channel in our niche, go to
16:36the video tab, and take a screenshot. We
16:38upload this screenshot to ChatGPT. We
16:40ask it to choose the most clickable
16:41thumbnails on the channel and explain
16:43why they work. What colors, how much
16:45text, what is in the image, where the
16:46eye is drawn. It selects the best ones
16:48and breaks them down in detail. Then we
16:50give it the topic of our video and ask
16:52it to create a thumbnail in a similar
16:53style. Not a copy, but our own based on
16:55the same principles. And now we have a
16:57thumbnail built on what is already
16:59working in this niche. There is a
17:00separate prompt for the video
17:01description itself. It’s also in the TG
17:03channel. We just copy it, open Claude,
17:05and paste it into the chat. Here you
17:07need to replace this text with your own
17:09keywords. To find them, we copy our
17:11video title and go back to VDQ. We click
17:13research, paste the title, and search. A
17:16bunch of keywords will pop up. We pick
17:17the ones with high search volume and low
17:20or medium competition. That way, the
17:21video has a better chance of being found
17:23faster. If you find a word with high
17:25competition but good volume, that's fine
17:27too. We choose at least five and copy
17:29them. We go back to Claude and paste
17:31them here. And in the script field, we
17:33just write "use the script above."
17:35Submit it, and the description is ready
17:37. Now, as promised, why did you set it
17:38to "unlisted" at the very beginning?
17:40When you upload a video this way, no one
17:42sees it, and YouTube can quietly check
17:44it for copyright issues and restrictions
17:46. If everything is clear, open it to the
17:48public. If a problem is found, you fix
17:49it, and the bad version doesn't get a
17:51chance to reach the recommendations. And
17:53the second secret about the American
17:54audience. Publish the video when it's
17:56evening in America. For Moscow time,
17:58this is night or early morning. The
17:59first few hours after release are the
18:01most important. If your viewers are
18:02asleep at this time, the video will have
18:04a slow start. Set a scheduled
18:06publication so you won't have to stay up
18:07at night. And keep one more thing in
18:09mind. If your video hasn't received any
18:12impressions or views, just relax. Don't
18:14delete it right away and don't give up
18:15on the channel. Keep posting, because
18:17new channels sometimes need time for the
18:19algorithm to notice them. Analyze your
18:21performance after at least 15-20 videos
18:24. And finally.
18:25If you want to work with me or my team
18:27to launch joint partner YouTube channels
18:28, I'll leave a form in the description.
18:31Fill it out. I will contact you
18:32personally. Also, I recommend watching
18:34this useful, step-by-step, free lesson.
18:36Good luck.

About this transcript

This page contains the full transcript of Как Я с Claude Сделал $32,740/месяц на ИИ YouTube Канале (пошагово) by Shukur Nabiev | YouTube, generated from the public captions YouTube serves with the video. The transcript has 4,321 words across 624 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.