Master Sora 2 AI For Faceless YouTube Videos (Complete Guide)

Sora2 just came out and it is incredible for faceless YouTube videos. But here's the thing, nobody knows how to actually use it to get the results they want on YouTube. So in the next 10 minutes, I'm gonna show you how you can unlock the full potential of Sora2 for making full YouTube videos while saving you tons of credits for the cheapest price possible. Step by step for making faceless YouTube videos. Let's get started. First things first, we need a way of actually accessing Sora 2. Right now, the ways that you can access Sora 2 are either through the app, which is available only for iPhone, but I do not recommend using that way because there's a watermark on the videos. Also, you can use the website if you're paying for the pro version of ChatGPT, but that's very expensive and there's also a watermark. So my recommendation is to use an online service like OpenArt, which is what we're going to be using for the purpose of this tutorial. OpenArt is basically an all-in-one AI platform that has tons of different models, not just Sora 2. So I'll leave a link down in the description if you'd like to try it out and follow along today. So once you're into OpenArt, you'll see the main dashboard right here, and we'll be focusing on generating videos with Sora 2. So we'll click video right here, and once the video generation page opens, we can go here to select a model and then select Sora 2. Now before we generate some videos and I show you how to use Sora 2 correctly in order to save tons of credits and save tons of money for making full YouTube videos and how to get better results with it, you might be asking, what types of videos can Sora2 even help me to create? Well, if you're running a faceless YouTube channel like I personally do, running a bunch of different channels in different niches, and I've gotten millions of subscribers across my channels, then you'll know that for these channels you need content. And whether you're running a health channel like Yellow Dude right here, which pulls in 5.5 million views per month and they just have an animated character doing different workouts, showing how to do them, or you're running a psychology channel like Brainy Dose right here, which has 2 million subscribers, or you're running a personal finance channel like Chris Invest, for instance. Or maybe you're even doing entertainment-based videos like this channel right here, which does videos on Uncharted Mysteries. And if we take a look at this video right here, what you'll notice is that this is literally an AI-generated background on this video. And then the whole rest of the video is literally just AI-generated images with a voiceover on top of it telling this scary, mysterious story. Of some terrifying prophecy. And this video got 700,000 views in just a couple months. And if you've ever run a YouTube channel like I have, you know that one of the hardest parts of running a channel is finding content for the channel. And historically, we would have to find content from YouTube itself, from stock footage websites like Storyblocks. But oftentimes we couldn't find the exact clip that we were looking for if we were using something like Storyblocks or if we were using YouTube for finding clips. But now Sora 2 makes it possible to generate literally any clip we could ever imagine. And even if we don't know which clip to put on our faceless YouTube videos, we can take a section of the script, put it into ChatGPT, and tell it to give us a prompt for generating Sora videos. But we need to make sure that we're generating the videos correctly, which I want to show you how to do right now. And this will save you tons of credits because make no mistake about it, Sora 2 is incredibly expensive at the moment, unfortunately. So if you are going to be generating full YouTube videos with Sora too, in my opinion, it's not really worth it unless you use the tactics that I'm about to share with you because it'll save you tons of credits. So when we jump into OpenArt and we take a look at this right here, you'll see that OpenArt allows us to do text to video in which we put in a text prompt and it generates a video. We can do image to video in which we put in an image and then we can animate it into a video. And then there's also elements to video, audio to video, video upscale, lip sync video, and motion sync video. Now for Sora 2, the only two options we can do are text to video or image to video. And let's start with text to video. As you can see, there is both a fast mode and a pro mode. The pro mode is going to be a bit slower and it's going to cost more credits, but the result will be a little bit better. For example, if we take a look at this video right here, This video was generated using the fast mode, and as you can see, it's still a pretty good video. But if we take a look at the pro mode video that I generated right here, it is a lot better. It's a lot sharper. The details are there a bit better, and it even changes camera angles during it, and it really utilized the prompt that we gave it to give a very accurate result. So my personal recommendation would be to use the pro mode for the absolute best result, and then also make sure that we're generated it in a 16:9 format since that's the format that we want for YouTube long-form videos. Now, did I mention that Sora 2 is expensive. Well, it is. And if we take a look at it right here, for a 12-second clip, it's going to cost us 1,800 credits, which based off of my $56 per month plan that I'm paying for OpenArt, where I get 24,000 credits a month, that comes out to over $4 just for a 12-second video, which is absurd. And we don't want to be spending that much because if we compare that to the fast mode for a 12-second video, it's only 600 credits, and the math on that comes out to only around a dollar for a 10-second video, which is still very expensive. So my personal recommendation would be to only use the Pro mode for generating the intro of the video, and the first 5 seconds of the intro are the most important when it comes to faceless YouTube videos. So I would recommend selecting 4 seconds, doing Pro mode, which would be about 600 credits, which would be about a dollar, which in my opinion is worth investing a dollar into the intro of your YouTube videos. But that being said, you don't want to use the Pro mode for the entirety of a 10-minute YouTube video because it's going to be super expensive. And I'll show you in a second how to generate the rest of your YouTube video for insanely cheap and how to do it generating up to 20-minute videos. But first of all, I want to show you how to prompt Sora2 correctly so you get the absolute best result from the generations that you're doing. And I came up with this framework to make it a bit easier for you called the Sora ID framework. And basically that's an acronym that stands for Subject, Objective, Realm, Atmosphere, Imaging, Details. Now you might be saying, Matt, what in the world does that even mean in the context of generating videos? Don't worry, it's actually very simple. Let me explain. And this is going to actually save you a lot of credits because you'll be able to generate exactly what you're looking for. So let's start with S, which stands for Subject. We want to be hyper-specific as to who the subject of our video is. For example, instead of just saying person, you want to say something like athletic woman in her late 20s with confident posture. Make sure to include subject's age, physical characteristics, and emotional state. Then for objective, we're going to describe the action that our character is taking. And of course you can be basic with it and say running, but if you want to be more descriptive, my recommendation would be to say something like jogging at a steady pace with rhythmic breathing. And then R is for realm, and this is basically just the environment and the setting. For this example, it would be a misty forest trail with dappled sunlight filtering through ancient oak trees. Feel free to include time of day, weather, and atmospheric conditions. Speaking of which, that brings us to A, which stands for atmosphere, which is basically the lighting of the image. And Sora 2 is incredibly good at lighting scenes. So we can say golden hour warmth with soft shadows, creating an inspiring, peaceful mood. And then I stands for imaging, which is basically the camera work and how we want the camera to be positioned and if we want it to move. So we can say, for example, tracking shot following from behind, gradually then moving to a side angle. And then D stands for any additional details that we want to give, like cinematic color grading, shallow depth of field so we can get a blurry background, and 4K photorealistic quality. So here's the entire prompt that we just put together using the SoraID framework. And then what we can do is paste that prompt into here. And what is kind of cool about OpenArt is that we can enhance our prompts directly onto here so we don't have to go to open up ChatGPT and ask it to make the prompt better. We can turn on Auto Enhance, and essentially what that's going to do is expand upon all the stuff that we gave it and make it a bit more in-depth. Now, this is a very in-depth prompt, so we really don't need this. The only reason I would recommend using this is if you have a very short prompt, like for example just a woman running, then you could turn Auto Enhance on. However, it does use an additional credit when you use it, which really isn't that much when we're spending 600 to generate a video. Then we can choose if we want sound on our video. I would recommend keeping that on, and then we'll keep Pro for model. And if we're generating an intro, let's say this is a health video about the benefits of jogging, then we just want to do 4 seconds because this is just the intro of the video cinematic. And then just a quick note as well, when you do have Auto Sound on, it is going to add an additional 10 credits onto it right there, just so you're aware. Then you click Create, and then that will be generating. And with OpenArt, you can generate multiple different videos at the same time depending on the plan that you're on, which is pretty cool. And while this video is generating right here, let me show you a really cool feature. So if you see a video that you really like on another another YouTube channel, how you can basically clone it then with AI. So if we really like the intro scene of this video right here of a giant tsunami crashing down in black and white, we can go ahead and we can type that out all manually. Or what we can do is actually take a screenshot of this video right here and then go back to OpenArt, go to Image to Prompt, and then you can actually upload that screenshot. And then once that image is uploaded, it's going to actually generate us a prompt based off of it. And it's actually creating that prompt right now. And boom, just like that, our prompt is done within a couple of seconds. And now we have an entire very in-depth prompt based off of what's going on in that image. It looks perfect. So let's just click Use Prompt. From there then I'm going to keep the Pro mode because let's say again that we are generating an intro for our Faceless YouTube video. And then I'm going to click Create. Now while that's generating, let's talk about image to video. So we can go to Image to Video right here and we can input any image and then generate a video with Sora too based on that image. And what's cool about OpenArt is that if we go to Image Generator right here, we can generate some images of whatever we want using all of the best models that are out there. The model I recommend using for generating images would be Nano Banana by Google. It is by far the best from what I've seen. And let's say we're running a cat channel like this channel right here called the Purring Journal, which has 200,000 subscribers, and we wanted to generate a video about The best tricks to teach your cat and a great way of keeping people's attention at the beginning of a video would be a cat doing something crazy. So let's go here to prompt and let's do a cat standing on a diving board in the Olympics. And then we can put in reference images if we want a particular cat. But for me, it really doesn't matter that much. And for output size, we're going to want to go widescreen since this is going to be a YouTube video. We'll do 2 images and click create. And these images right here, they're okay. They look a little bit animated. Honestly, I want it to look a little bit more realistic. So we'll add a little bit to the prompt saying photorealistic as if it's being shot on a TV camera in 4K. There we go. And we'll do a few more images here. We'll do 3 now just so we can have some more options. And now these are looking a lot better. Oh, wow. These are looking really good. So let's just go with the best one, which in my opinion is this one right here. Then we can click "Use Image," go to "Image to Video," and then select Sora2. And then we can tell Sora2 what action to put on this video. So we can tell it, for instance, "The cat jumps up and down on the diving board before doing a perfect flip off of it and into the water. The camera follows the cat." and we don't have to go as in-depth on the prompt due to the fact that we already have the image generated for this one. We're really just describing the action that we want to happen. And I want to select Pro so we have the absolute best quality. And I want this to be an 8-second video right here. 1,200 credits. Wow, that's a lot. But remember, this is just for the intro for our cat channel. This does not need to be the entire video. In fact, we can use stock footage and other methods that I'm going to be getting into in a second for generating the rest of our video. In a matter of seconds for a 20-minute video? Yes, it's actually possible. I'll show you in a second. But first, let's generate this incredible intro video right now. Now let's take a look at what Sora 2 generated for us for this tsunami video right here. As we can see, it looks pretty darn good just for the intro of our video. And again, it's keeping it purposefully not ultra realistic because that's the prompt that we gave it. If we, for example, wanted to give it a photorealistic prompt and we didn't say a black and white illustration, it would obviously be a lot different. And boom, now one of our cat videos is done. So let's take a look at it right here. Wow, that's honestly pretty dang good. And that is much more likely to get attention in the beginning of your video when you compare it against a stock footage clip or just some generic Google image. And the most important part of a faceless YouTube video is the beginning of the video. And I know a common question I'm going to get is, how do I keep a consistent character throughout my Sora 2 generations? Well, the best way to do that would be to go into OpenArt, go to Character right here, and then you can upload 4 different images of a character that you have or one that you want to create. My recommendation would be to use Nanobanana to generate 4 different images of a character, one from the front, one from the side, one from a diagonal point of view, and then one from the opposite diagonal point of view. That way you can upload them all onto here, create a character, and then you can go to Image and then you can generate images based off your character. You can even pose your character and everything else. Like, for example, if we wanted to have this cat be like the host of our YouTube channel, then we would want it to look the same. Then we would generate images based off of this character in the different environments that we want them to be in. And then when it comes to generating actual videos with Sora 2, we would go to video and then we would upload those images that we generated from our character, and then we can generate consistent characters with Sora 2. Now unfortunately, lip-synced videos is not yet available with Sora 2 for specific things that we might want our character to say. So if we do want to go to lip-syncing video, as you can do with other video models, then you can choose any of these models right here. And both OpenArt Lip Sync and Kling are both very good. So if you do want a character that can actually talk and everything, that would be the way to do that. And then you would save Sora for doing things like the intro, for example. And now to give you as much value as I possibly can, I wanted to generate a a full YouTube video using AI so you can see exactly how to do it completely faceless and also using Sora too. So the way that I would go about doing this is, first of all, we would need a video idea. Now, you might already have video ideas, so you can just use the ideas that you have. But if you don't have an idea, then you can jump into a tool that I actually made called TubeMagic. And I made this specifically for myself for helping me run channels, but I do have it public so anybody can use it if they want. And we have a video ideas tool right here. And what we can do is take the name of a channel like the Purring Journal, for example, this cat channel, put it into our video ideas tool, click Generate, and then TubeMagic is going to generate us video ideas based on what has already performed well on this channel and then also indexing it across other videos that have worked on YouTube. So that's a great title right there. 11 Curious Cat Myths Debunked. Prepare to be amazed. And that's a pretty good title. But let's generate another couple ideas right here just so we can choose the absolute best one. That is likely to get the most amount of views. 10 Ways Cats Secretly Help You Live Longer. That's also pretty darn good, honestly. But this idea right here is perfect. 10 Things Cats Hate. And it's actually proven to already do good on YouTube, as evidenced by this video that got 4.3 million views. So what we're going to do is take that title, go back to TubeMagic, go to script, paste in our video script right here. And this is a real reason that I built TubeMagic was because using tools like ChatGPT, for example, to write YouTube scripts it's going to index all of the internet, plus it's going to not be written in the best way because ChatGPT doesn't do as natural of writing as something like Claude, which we have built into TubeMagic. And then what's really cool is we have all of these viral hooks, intro frameworks, strategic ordering techniques, emotional way techniques, psychology formulas, all designed to keep people on your video as long as possible. And all of those will be selected by default. And then for context, we can go into videos like this one right here that already got a lot of views, paste in that link. We can also search on YouTube to find other titles that have gotten a lot of views, just like this video that got 6.6 million views. We'll copy that link and then we'll also add it as an additional context. And my recommendation would be to add 3 to 5 videos as context right here. But for the purpose of this, then we're just going to select our word count. Now we want to make at least a 10-minute video, so we're gonna select 1,500 words, which should be at least about a 12-minute video. And then we're using Claude for the backend, and then we just click enter. And this is going to pull the information from those videos that have already gone super viral and then write it using all of the methods built into TubeMagic to give us the absolute best script that we can possibly get. Plus, it'll be a more accurate word count because if you just use something like ChatGPT or Claude, it's never accurate with the word counts. And just like that, our script is done. 1,356 words. Perfect. It's an incredible script. So then what we're going to do is then copy this script right here. And this next part is the secret for generating full videos up to 20 minutes long in literally just a couple of seconds and then incorporating Sora AI into those videos. And the secret is my other tool called vid.ai. Now vid.ai can generate full videos up to 20 minutes long. And what's incredible about this is that if you go to create long form videos right here, what you can do is first of all give your project a title. We'll just put in the title of our video, 10 Things Cats Hate. Then we'll go to script and we'll paste in our script right here. And the reason we generated the script with TubeMagic is because there's all of those things that go into making it an incredible script. Whereas if we just use VidAI script generator, it's not nearly as good. Then we can select a voice. I'm just going to use my voice as default right there. And as you can see, our video will be 12 minutes long. We can choose any of these voices, by the way, but we're just going to do this one just for the purpose of this example. And the thing about AI is that it's a lot less credits, which is going to be a lot cheaper for generating this video that's 12 minutes long, as opposed to just generating the entire video with Sora. And I'll explain to you how we're going to incorporate Sora into this in just a second. But once we click Generate Video, then we simply select Next from this right here. And now our actual full video is generating. And this is super fast, by the way. If you see right here, it's already at 15% and the percentages are going very fast. That's due to the fact that we generate multiple parts of the video at the same time using a proprietary method, which makes it a lot faster than other video generation software out there. And just like that, our video is generated. And if we take a look at it right here, it looks like this. And as you can see, it's great. It's got captions. It's over a 10-minute video. However, there's one major issue with it. There's no Sora 2. So how can we incorporate Sora 2 into this video and benefit from the fact that we barely spent any credits on this when you compare it to a fully AI-generated video? Well, what we want to to do is generate an intro for this. So it's talking about why a cat runs away from a vacuum cleaner. So let's go back to OpenArt. And by the way, this video finished generating right here as well. And this is of a more realistic giant wave. As you can see, it's super realistic compared to this video that we generated. I just want to show you an example of that. So let's generate an intro for this video. So let's do a cat doing a backflip, jumping out of the way of a woman. Vacuuming her house, doing a backflip, scared, jumping out of the way. Boom, there we go. And let's do a nice 4-second clip right there and click Create. Now we got that generating right there. And while those are generating, we can jump back over to VidAI and we can adjust any of the stuff with the captions. In fact, we can turn the captions off if we want to, if we just want to have videos in the background. We can change them to any font that we'd like. We could put a border around the captions and we can change anything. We can also put music in the background of our video if want. And it's all royalty-free music built into it. And we can even change the captions if we want to. And what we can also do is go to Media, and we can also change this. Like, let's say for example we wanted a different stock video of a cat running right here. We can search that and then input any of these videos in the place of the video that it automatically chose based off of what was being said in the video. However, none of these are going to be as good as Sora 2, because with Sora 2 we can generate something that is exactly what we want to see. And there we go. Now we have an intro done. And that is amazing. As you can see, this is a video that we could have never gotten from any stock footage based platform because they're not going to have a cat doing a backflip trying to get away from a vacuum cleaner like we have right here. So this is absolutely perfect. Then what we're going to want to do is go ahead and download this video. We can also upscale the video if we want to as well, if we want it to be even higher quality. but I think this is more than good enough. So we're just going to go download right here. Make sure to select remove watermark and then choose MP4. And then with VidAI, we're going to go to the media tab and we're going to go to the first clip right here, the first 4 seconds, and then we're going to click upload. And then we can upload a file that's up to 150 megabytes right there. And our video that we generated is not even 1 megabyte, so it's completely fine. And now we got our video uploaded right here and we can select when we want the clip to play, because if our clip is longer than the time that we have allocated to the existing clip, then we can change it if we want. We're just going to keep it like that. Click Set. And now check that out. We have that as our intro. So now if we play this video, it's perfect. It literally corresponds to exactly what is being said in the video. And then throughout the whole rest of the video, we are just using stock footage. And if we want to go the extra mile on top of that, we can go back to Sora 2, and then we can just use Sora 2 Fast. Because it'll be a lot cheaper, and put those occasionally throughout the video. But it doesn't need to be throughout the entire video because we have stock footage for the rest of it. So I would just put it every now and then throughout your faceless video, and the rest of it can be stock footage just like that. When you're ready to export, you just click Export right here. It'll export your video, and then you can go ahead and save it onto your computer, and it's ready to upload to YouTube. And now if you want a complete tutorial on how I run faceless YouTube channels and I've grown a huge business out of doing it, you can check out this video next to me right here. Which will show you exactly how I run these channels step by step in a complete tutorial. And this video right here is essentially a free course. My name is Matt Parr. If this is our first time meeting, and hopefully I'll see you there.