[00:02] videos. And a few months ago, they didn't even [music] exist. But there's no secret formula behind it. Just some winning topics and good thumbnails you can recreate. But the worst part is you've only got two months left. Hi, I'm [00:14] everything from picking a topic to thumbnails with no team [music] and without showing your face. And I'll be using the smartest model right now, which makes all of this so much easier, [00:26] Claude Cable 5.2. But first, let's look at why right now is the best [music] time to start a faceless channel and whether YouTube is actually demonetizing them. I've watched dozens of videos on this and there are a couple of huge [00:38] this and there are a couple of huge details nobody [music] mentions. a huge range of styles. [music] Historical POV's, animated explainers, mini documentaries, AI avatars, gaming, ASMR, [music] [00:52] views and bringing in subscribers for the past 10 years. That's a good sign And where there's attention, there's money. I mean, you can see it. These are the stats for just 28 [music] days from [01:06] a channel that's been posting only one video a week since July. That's more than the average monthly salary in the US. But this channel earns that much That's education and science documentaries, one of the highest paying [01:18] different rates [music] to reach different audiences, so the same number than another. [music] To compare that, we'll look at how much a channel earns per thousand views. That's called [music] RPM. This [01:31] channel's RPM is $4.50 per thousand views, when the usual is $1.50 to $2. That means this single video with 1.6 million views pulled in roughly $7,000 [01:43] from just [music] one video. Looks like we've picked our niche. But before we There are plenty of styles to choose from, remember? But we need [music] one that grabs attention. For that, I go to [01:55] specifically for creating faceless [music] and topics get the most views and likes. That's just a signal for what visually grabs people. And as you can see, what if videos and explainers get [02:09] That's what we need because [music] it means high retention. And the longer people watch a video, the more the algorithm pushes it in recommendations retention also comes down to the visuals and script. In Content Factory, you can [02:23] just recreate any visual style without building it from scratch or you can just upload your own reference. For this video, I'm going with realism. If you experiment. This channel [music] uses simple stick figure animations and makes [02:36] realistic one from the beginning. [music] And notice the on-screen text explainers and faceless videos don't have any, even though they're a huge and keeps them watching. Okay, so we've got the niche, the format, and the [02:51] style. The next question is, which format is better for monetization? Long form [music] or shorts? are a little trickier to make money from, but they're a great starting point [03:05] want to watch your content. But if we want to earn more, we need longer videos. The longer the video, the more ads YouTube can show and the more YouTube [music] earns and the more it pays us. So I recommend aiming for at [03:18] we'll make five in a long form videos and [music] 30-second shorts. Creators were already making money from both shorts and long form videos before the AI content boom. Then [music] AI made these channels much easier to produce [03:31] bulk. And that why YouTube did cut off monetization for some of those [music] channels. But that doesn't mean every channel using AI lost ability to earn. they're posting. Here are the kinds of content that gets you demonetized. Mass [03:45] produced content that [music] offers zero educational value, emotional bait made for clicks, and [music] AI versions of real people giving advice on health, money, or legal issues. And people who actually read the rules [music] realized [03:58] competition, leaving the niche wide open. And that's the main reason I started this experiment. [music] YouTube officially allows AI content. The AI label doesn't hurt your reach or your ad revenue. So now that we [music] know [04:10] what YouTube allows, let's move to the next stage. if topic, writing the script, and generating a video that keeps your to research which topics [music] hit over a million views for different [04:25] creators again and again over the last five years and explore Content Factory to find three most viewed and liked topics. I also wanted to check what's platforms, especially [music] what's going viral right now, and give me the [04:38] can see how much [music] smarter Fable 5.5 is. I like this one. Let's go with it. And by the way, here's a pro tip for picking a topic. If a video gets over a thousand times [music] as many views as the channel has subscribers, that's a [04:52] sign there's demand for it. Look at this channel. [music] 3,000 subscribers, but its video got 5.1 million views in just 12 days. That's what I'm talking about. don't have to be a scriptwriter or study the topic inside out. Every [music] week [05:06] and I know the exact second viewers leave. I've tested different styles and scripts myself and put all of it into one skill for faceless video scripts. So I'm giving Claude the skill and asking him to write a script [music] on our [05:19] prompts in the description so you can recreate it too. In the [music] prompt, I asked for a five-minute video script where the shots change every five to 10 it feels dynamic. After that, it switches to 30-second shots. Fable takes [05:34] my structure, [music] hook, intro, body, recap, and outro. For this format, the hook usually starts with [music] what if. And the skill niche, breaks down the viral patterns behind the top-performing videos, [05:49] [music] and scans comments under popular faceless videos to find out what people does it start writing. Please don't repeat my mistakes. I once asked Claude to write me a script without giving it much detail and ended up [music] wasting [06:02] I wasn't even happy with. So it's better to use prompts that are proven to work need. Now, when the script is done, always give it a quick read and fix actually okay with what the script gave us, so I'll just keep this version. Now, [06:18] [music] video, I go back to HeyGen Content Factory. Now, I'm connecting second skill. So, this one writes prompts that handle the camera moves, the lighting, and the storyboard so you get far more usable [06:32] [music] So, I'm asking Claude to turn a script into a 5-minute video in a realistic style, add motion graphics and captions explained, generate music and sound effects, and edit everything together. [06:45] And HeyGen here can handle the voiceover too. [music] You know how most videos like this use that typical AI sound that turns people off right away and that can choose [music] a more natural sounding voice through Eleven Labs or Seed Audio [06:58] in HeyGen MCP and give it a specific direction which makes your content more and the mood. And the instruction can get much more specific. How fast the narrator speaks, where do they pause, which words they emphasize, and how the [07:11] narration lines up with [music] the visuals. In fact, we measured what winning narration sounds like. Our team broke down 10,000 scripts and 1,000 faceless videos tracking narration down to 20 [music] milliseconds and built [07:23] to the difference. >> What if a pophys hit Earth on April 13th, 2029, 100 km away, the light arrives first? [07:36] >> What if a pophys hit Earth on April 13th, 2029, 100 km away, the light arrives first? >> And to make your content even more recognizable character. It's a proven move. This account, for example, started [07:51] posting videos with a gorilla. But I would say that generating the video is title and thumbnail, it'll get [music] maybe 1,000 views if you're lucky. So, a those and give you the guides that you can use. Now that the video is done, [08:06] >> About 56 km >> the rock starts [music] to crack apart, but it's too big and too fast to break up in time. From the first glow to the ground, about 10 seconds. On impact, rock vaporizes. [music] [08:19] A fireball several kilometers wide rises over the plains. In under a minute, a crater opens, about 4 and 1/2 km across, nearly half a kilometer deep. Debris rains down across [music] the region. From orbit, it's a single point of light [08:33] and a plume of dust that keeps growing. >> I [music] actually keep watching this. minutes, but you can watch the full video through the link in the description. But look right here. See this shot? It came out as a bit of slop. [08:50] you have to re-render the entire 5 minutes [music] when this happens. Just tell Claude which shot needs fixing and what you want changed. Then ask it to the updated video. And because Fable keeps the whole project in its head, [09:04] twice. It already knows which prompts made that shot and where it [music] sits made that shot and where it [music] sits in the edit. [09:17] this skill matters. Better prompts mean fewer shots to redo, and the model matters, too. In my tests, [music] Claude 3.5 gives way less slop, but you we've got our first video. That's a good start, but we need more if we want to [09:31] build a channel and make [music] money from it. So now, let's turn this into a I'm just going to [music] write, make two more 5-minute videos on the other Claude does deep research for each topic individually, every video [music] gets a [09:45] real upgrade you need to know about in the new Fable. [music] Then we'll turn you how to use those on Instagram as well. [music] While it generates, let's with thumbnails. YouTube has a built-in A/B test feature. You upload three [10:00] [music] tests them on real viewers to find the winner. Look at these three thumbnails. Which one would you actually click on? This one grabs attention, tell us which one will win. [music] The winning thumbnails is the one people [10:12] click and actually stay to watch. So to create strong options for the test, I First, the thumbnail has to look like a moment frozen in [music] action. Second, keep the text to four or five words max. We have less than a second to grab [10:27] it has to match what the video is actually about, but still have a click on the thumbnail. I put these rules into a thumbnail guide, give it to prompt to create three thumbnails and title options for our video. So I kick [10:41] some MCP, starts generating them, [music] the title options, descriptions, tags, everything you need. All that's left is to upload the video, add the thumbnail and details, and select yes under [music] AI use in YouTube Studio. [10:54] Now to start making money from your account, you need a total of 1,000 and 4,000 hours of watch time. Even if just one 5-minute video gets 100,000 [music] views and people only watch half of it, you've already hit your 4,000 [11:08] took me just [music] 10 minutes to generate. We didn't do a single thing by hand. AI handled every step. And that's exactly how it should be [music] used. All I gave it was four prompts, one to find the topic, one to write the script, [11:22] one to generate the full video, and [music] one to direct the voice. That's videos we started earlier, [music] they're ready too. >> What if the oceans vanished, leaving ships and sea life stranded? [11:35] This impossible thought experiment removes seawater and its salts without evaporation. The ship loses buoyancy. Gravity remains. Here, the bottom lies close. [11:49] Freshwater, land ice, and the air above us remain at first. >> Now let's make shorts. They're cheaper to produce [music] and just one that to your channel. I'll show you two ways to make them. The first way, ask the MCP [12:02] to cut three shorts from our video, 9 [music] by 16, 15 to 30 seconds, each with a strong hook. >> In our city, 100 km away, [music] the flash is blinding, but too distant to burn. 20 seconds later, the ground [12:15] shakes. And almost four minutes after the flash, the air blast arrives. [music] Was across the city shattering Closer in. In about 20 km, almost nothing is left [12:27] >> The second way, generate brand new ones inside [music] the MTP. And if something feels off, give Claude exact feedback. Don't say make this [music] more interesting. That doesn't mean anything. Say something like the opening sentence [12:39] asteroid question, cut straight to the impact, and keep the explanation in the [music] videos with different hooks and topics to see what the algorithm actually picks up, which means you need topic research. And yeah, sometimes [12:53] research looks like scrolling Reels. Okay, back to testing. Remember Content visual styles? It also creates what if videos [music] from scratch and we can use it through Pixel MCP. So here's the next prompt. Let's take a [13:07] So here's the next prompt. Let's take a look at what Pixel made. from space? A giant halo from the equator, almost a line. [13:21] A thin band of rock lies around the equator and you're looking along its edge. Move to 35° north. [13:33] Now the same ring opens across the southern sky. The ring never changed shape. You changed where you stood. Remember I said Fable is much smarter? [music] That means you can give it tasks [13:48] that sound impossible, like generating 100 Shorts at once, and it will handle exact same Shorts and post them to Instagram. Now, how do we monetize them? For Shorts ad revenue, [music] the bar is 1,000 subscribers and 10 million [14:01] valid Shorts views in 90 days. On Instagram, monetization works a bit differently. [music] Instead of waiting for YouTube ad revenue, brands just pay companies will literally slide into your DMs and ask you to feature their product [14:15] and it works. [music] When you find a format that consistently hits, just Content [music] Factory and analyzes the pacing and replicates the format for your next video. Just remember the golden rule. No fake [music] AI doctors, [14:29] lawyers, or financial advisors giving advice, and no AI replicas of real Keep it creative, keep it educational, and play the long game. And by the way, beginning, [music] not all of the videos go viral. You can check it yourself. [14:41] Almost half of the videos got under [music] 100,000 views and only four crossed the million. But the biggest one alone made about $22,000. [music] So one hit pays for all the misses 20 times over. So judge your [14:54] channel by 10 videos, not one. [music] And don't give up if your first video doesn't take off. That's normal. Even one of those unsuccessful videos with 100,000 views [music] brings the channel about $450 bucks. Post one video a week, [15:06] and even if each gets 100,000 views, that's almost $2,000 a month. [music] exactly why Faceless is so popular. Faceless Instagram accounts and YouTube limits here. You can experiment, get creative, and build a real side hustle [15:21] started. The challenge is complete. Now I'm passing it over to you. Pick a topic around, [music] create your account, and upload your ChatGPT combo. I'm going to leave the MCP link along with all the skills and [15:35] prompts [music] and guides we used in the description. And as always, if you the like button, subscribe, and I'll see you guys in the next one.