Transcription
Here's how to create your first AI-generated song with a consistent character in a full music video. Right, so watch this.
This entire song is made with AI. The vocals, the music, the visuals, everything. And if you look closely, the character follows the lyrics and stays consistent for 90% of the clip. You can check out the full version by clicking the pop-up over here. And it's full HD, too, which is wild.
Now, here's what we're going to do today. We'll create our own AI-generated song. I'll show you the easiest way to make visuals for your track. We'll go over some advanced workflows to get more personalized results. And finally, I'll show you how to insert yourself into the music video using AI. Let's get started, friend.
Step number one, find your inspiration. Like any creative process, I reiterate, we have to find our reference. What, what do you want your song to be about? How do you want your song to sound like? Do you have a favorite artist? For me, and for the purpose of this video, I'm going for something that's calm, instrumental. I am obsessed with cinematic, emotional music, something that makes you feel like you're in a dream or some sort of a memory. And I'm thinking of "Saturn" by Sleeping at Last. Don't judge me because I do love sad songs. I do love soul-soothing music, okay? It's my thing.
And while I'm thinking this, I'm asking myself, why am I doing this with AI even? Well, I'm not a professional musician, right? But I do have taste. I have a vision. I know what's good. I know what something should feel like. And I believe that AI bridges that gap, right? It gives me the ability to create without limits. So, the very first thing that I'm going to do right now is I'm going to go to YouTube and grab the link of the track that, that I like that I want to download. In this case, it's going to be "Saturn" by Sleeping at Last. That's our source, not image, but our source audio.
Step number two, we need to prepare that audio so we avoid copyright to generate the audio track. We're using Suno, which is one of the best AI music platforms out there, but, but it will not let you use copyrighted tracks, and that's how it should be. But since we're using this as an inspo, we're not trying to copy. We're trying to get inspired. Here's a little trick for you. We're going to go over to vocal remover and we're going to upload our MP3 file. So, the music that we downloaded and we're going to do three things. One, we're going to change the pitch. So, we're going to increase the pitch to plus one over here, plus one. And we're going to slow down the speed to 80%. And we're going to click save. The reason why we're doing this is that AI is not able to match the, the original track to the copyrighted material because we are only using this to get inspired. Okay? We know the intentions here.
Then what I would do is that I would definitely separate the vocals from the instrumental. Why? Well, again, we want the feeling of the song. We do not want the original lyrics or, or, or the melody, right? We just want the feel. So, we're going to head over to the remover tab on the same side. We're going to drag our modified file that we adjusted, and we're just going to let AI separate the vocals from the instrumentals. We're going to download the version with vocals turned off all the way down. This will give us a, a clean backtrack that, that will not trigger any copyright filters. So, we're going to head over to save over here and just, just download the instrumental track.
Step number three, we're going to upload our modified audio to Suno, right? The track, the one that doesn't have any vocals. We're going to head over to audio. We're going to click upload and we're going to select our modified track over here. And we're going to click cover and wait for a few seconds. After we have uploaded our track, we have two options. We either generate instrumentals, uh, and we build up from there ourselves, or we can directly immediately add our own lyrics and let Suno compose the vocals from it. I usually do a bit of both, right? I don't, I don't like to do the latter because if I'm just going to add lyrics to an song that's copyrighted, it's just going to give me something that's extremely similar. We want to avoid that. We're not here to steal. We are here to get inspired.
So, what I'm, I usually do, and especially if I do not have the lyrics figured out, I'll create placeholders. I'll, I'll go into the lyrics section and I'll add placeholders like "oh" and "hmm," humming sounds. I know it sounds a bit silly, but by, by humming, it's going to give you space to explore the mood of the music first, then after you can lock in the, the, the words, right? So, that's what I'm going to do right now. Gonna just paste in some hums. Um, I don't even know if that, that is that how you call it in English. I'm not native. So, yeah, we paste it there. Um, and we're going to add our style influence. Right now, this is where it gets a bit spicy. In the style section, we're going to tell Suno exactly how we want our track to, to sound like, right? So, it's sourcing our audio track and we have to tell it, okay, this audio track that you have, that we, that I want you to get inspired from, I want you to have it in, I don't know, like in, in style of a pop music or like a chill music or emotional guitar or piano, etc. This is where most of your effort has to be put on. If you want a song that sounds like, like you, or like maybe like you want a song that sounds like your favorite artist, you have to tell AI, well, you have to tell AI that it has to generate this song or restyle this song or create a cover of the original track in this particular style.
And what I do normally is that I take Suno's auto-generated description over here that, um, generated when we uploaded our modified audio or track. I'm going to copy this and I'm going to feed it to ChatGPT and paste in this prompt. So, we're going to head over to ChatGPT and we're going to say, "Describe the music style of." You can add in yourself, or maybe you can add in an artist that you like, in four sentences using this format. And now here you're going to paste in Suno's auto-generated style prompt and hit generate. This is what Chat comes up with. We're going to copy this and we're going to paste the refined description into Suno's style input section over here and we're going to click generate. Let's see what it comes up with.
[Music]
Okay. Well, we've got something that's very close to our reference. Still a bit unique, though. It's very familiar. It's definitely hitting copyright strikes. This is not great. This is almost like you're stealing or like creating something that's extremely similar. That's not what we're trying to do here. Your AI creation has to feel unique. I, I do not endorse stealing or creating things that are extremely similar to, to your, your source reference. You have to be creative. You have all the tools. Now, just do the extra work. And for you to do that, we have to adjust the sliders accordingly until we, we achieve our desirable results. So, we have to adjust our audio influence or style influence so that we get things that are close enough but not exactly the same and they do not remind us of the original track. If it reminds you of the original track, you're doing something wrong and that's copyright over there.
Now, step number five, ready to add the lyrics. We have our melody locked in, right? Let's listen to it one more time. I like it. But it still sounds very, very familiar to the original track. But what we're going to do right now is that we're going to add some lyrics to change that. And what I do to make sure that we stay away as much as we can from the original track is that I would create a cover of this cover and paste in my lyrics while making sure that I adjust the settings accordingly. But a quick note, you cannot just paste a paragraph of lyrics and hope that AI will get it right. You need to structure it correctly. So if you add your lyrics, for example, if you take a look at my lyrics, I've clearly labeled each section. We have a verse, we have a chorus, we have a bridge, we have an instrumental, and what you have to do is that you have to add all these tags. I, I'll call them tags before every section in brackets. That's going to help AI to understand, uh, the, the structure of your lyrics and build something coherent. And I would also lower the audio influence to 10% so that we can focus more on the lyrics than the reference melody because again, we need to stay as far away as possible from copyrighted music.
But is this cheating? I know you just asked that. I keep asking that to myself. Am I even a musician if I didn't play an instrument or I didn't sing a note? Well, for me, I do play instruments. I do sing. I do write poems. However, I'm not a producer. But then when I think about it, I'm making all these creative decisions every step of the way. So, I'm directing the feeling, the, the vibe, the pacing, and that's what a producer does. Um, AI just makes it possible for me and more people to become producers. And honestly, to me, that's very exciting because I do have five songs that I have written in the last few years that are ready to be produced with AI. And I'm extremely excited about it because who in the world knows better than I do? Me, right? And that applies to you, too. I don't need another producer to tell me what to do. Now, I can be my own producer. And now I can determine what my music should sound like and should feel like because, you know, I trust my gut feeling.
Okay, so this is what we get from Suno. Let's listen to it.
>> I could dream [Music] with eyes open and everything. [Music] Swimming in waves, drowning in tears, walking in nightmares. You found my fear. [Music]
Look, this is good. This is good. However, it needs a lot of work. The lyrics is not optimized whatsoever for the track. It's way too long. It doesn't really make sense. It was a draft, a draft from one, one of my very old songs. But for the next part of this tutorial, I'll be using a song I generated a few days ago by following this exact steps.
Oh, the stars don't shine the same. Oh, the night just calls your name. Oh, my heart's a broken charge. Tracing the lines where we fell apart.
But this time I did take my time to optimize the output and I will be releasing the song soon, sang by me. In this case, because this version that, that you're listening to is generated with Suno and all the visuals will be AI generated. So make sure to turn the notifications on for that and give it a listen. And I'm going to apply the exact same workflow that I'm showing you today for my final result. Cool. We have our track. It's ready. Let's generate some visuals for it. Listen to it one more time.
>> Shine the Oh, the night just calls your name. Oh, my heart's a broken heart. tracing the lines where we fell apart. >>
There are two main methods for generating visuals for your track. First is that a quick and easy and efficient way, like think, think lyric videos, eye-catching visuals, lip-syncing, and short storytelling. It's perfect for social media, but also great for YouTube. This method though does not require much technical skills, and it's ideal for people who are just starting out with AI enthusiasts or anyone just looking to move extremely fast. As I've said before, in the AI space, speed is more important than quality.
And then we have the advanced workflow, which is more time-consuming, but it gives you higher quality results and more creative control. However, it does take time and it's for people who want to go in and push things to the next level. But first, let's start with our first option. Open Art just released a new feature that's called Music Video, which lets you quickly generate 10 to 30 second videos in both, uh, 16 by 9 and 9 by 16 formats. Again, YouTube, TikTok, short form. And there's three modes like you see over here. We have style, we have story, and we have sing. And we'll go through all of them three, but I have already my favorite one. The very first thing that we need to do is that we're going to upload our track here. And mine is about 30 seconds long. So this is the chorus. Again, keep in mind this is a short form chorus. You need the thing that hits. I'm selecting the 16 by 9 aspect ratio. And I'm choosing my mode, which in this case is going to be the style mode. What does this mean? This is the mode that you can customize your video's aesthetic. Right? You can either pick from an existing template or you can describe and create your own. And for this track, I really want to go with the psychedelic visual style. I think it fits the vibe really well. And this style mode is, is great when you just want to have beautiful abstract visuals without a specific narrative that, that has to have this continuity. And what I think that this mode works best with would be with lyric videos, right? You can add lyric videos on top of it and you can have a, a readymade production to post on social media.
So now we're going to create our story. We're going to hit create our story and the AI will start building something for you. It's going to start analyzing the lyric and it's going to start, uh, generating images, um, to match the lyric and finally it's going to compile everything into a video. It's going to generate a video from those frames and this will take about 3 to 5 minutes, right? It's not very long. So, let's wait for that.
Okay, here's what we got. Let me play it for you.
Oh, the stars don't shine the same. Oh, the night just calls your name. [Music] Oh, my heart's a broken shing where we fell.
This looks pretty dope. I think if we add lyrics on top that this, this will be great and ready to post on socials. And this is exactly what, what makes this mode, I think, amazing for quick content creation or short form content because you don't need to to worry about consistency. No, this is like an abstract, very cool visual, uh, production that, that you would not be able to do otherwise or that would require you like extensive knowledge in, in graphic motion design and all of that. Now you just do it with a click of a button. Very convenient. Um, I'm going to move now into the story mode. And this one is ideal for narrative-based or, or social-friendly videos that, that feature some sort of a consistent character. And that character can be you, or that character can be an AI avatar, or, or, or a cartoon that does not exist. And you can either select auto story. It's going to create the story for you, or you can select one star story. And what's cool about this is that you can have your own consistent character in the video, right? And we will be experimenting more with that in the next mode. But for now, what I'm going to do is I'm going to pick one of the built-in avatars just for this demo. I'm going to describe the style as romantic. Actually, you don't have to, right? You can just hit generate. Let the AI do its job.
Here's what we get.
The stars don't shine the same. Oh, the night just calls your name. [Music] Oh, my heart's a broken shing where we fell.
Honestly, this is pretty cool. Again, I'm not expecting studio quality grade. Like, don't comment me at the comment section. Oh my god, Serio, this is not, it's not meant to be. It takes 5 minutes. It should not be, right? It takes 5 minutes. For what it is, though, I think that it's quite impressive, especially if you want to do things quick. And the thing that I really like is that our character stays consistent throughout the video, which is quite hard to achieve and would require you very complicated coffee UI workflows to achieve this result, especially in 5 minutes. And that alone makes it feel more polished than most AI-generated stuff. Again, it's not perfect, but it does the job. You have to put a lot of effort. Uh, and I always keep saying that if you just expect AI to do it for you, you're not using it right. And it does take editing and polishing and, and, and curating. There's an editor built in that, that lets you fine-tune your video and adjust the scenes, replace frames. Um, I'm not going to dive into that today, but now let's talk about mode number three, which would be my personal favorite, which is sync mode.
Now, this mode lets you lip-sync your audio to an avatar. You can either choose one of the pre-trained characters or you can upload your own image. In this case, I'm going to upload my own photo because I'm very curious to see how my AI will represent me and how, uh, are they going to be singing. But one important thing that I have to note here is that if your audio isn't clear, and I feel that my audio is not that clear, the vocals are already in the mix and, and would be very hard for the AI to detect the words and of course the lip-syncing will be off. So my trick would be that I'm going to upload for this case an acapella version of my track with the vocals only. This will give the AI a much better understanding to, to accurately match my mouth movements because it's going to understand the lyric.
Going to upload my audio. I'm going to upload my source image over here and we're just going to generate it and see how it does it with my face. My pretty little face. And as you're seeing right now, I'm not wearing a hat anymore because my hair transplant is hair transplanting. Um, feel more confident about this. But, um, yeah, let's see what it comes up with.
Oh, the stars don't shine the same. Oh, the night just calls your name. [Music] Oh, my heart's a broken heart.
Okay, so this actually does look like me. And the, the lip-sync is surprisingly good. And we're currently using Open Art's lip-syncing model. You can choose Hedra or other alternatives if you wish to achieve more quality results. But honestly, I think this is very cool, especially if you're just getting started with AI. Barrier to entry is super low for you to start creating and understanding how these tools work because future of content, my friend, is AI. Get used to this fact. If you want to try it out, you can head over to Open Art. I've actually dropped a link in description and you can play around with it. See what, what you can create. And most importantly, have fun while doing it and let me know what you come up with.
All right, so listen up. If you are not just an AI enthusiast and you're not messing around for fun, and you are someone who wants to take AI content production to the next level, pay attention because the rest of this tutorial is for creatives who have time, have budget, and want to achieve professional level results. The very first part of this workflow, which is using SUNO to generate the, the track, still applies because we have our music. Now, it's time to create photorealistic results with more control over our visuals. And to do that, we'll head over to Kora, which is hands down my favorite and in my opinion is the best photorealistic AI image generator for high-end production. Okay, high-end production, we're talking about the real thing here. And you can access it inside Enhancer AI. And once you're in, um, you will see this interface, clean interface with, with images and prompts that you can either use directly, copy the prompts from those images, or customize your own.
The very first thing that we are going to do is that we're going to select our model, not our model as in our, our fashion model, our actor, but our image model. We have two options over here. We have Kora Dreams for surreal, dreamy visuals, and Kora Cinema, which is for cinematic, moody results, more realistic. But for this tutorial, I'm choosing Kora Dreams because I want to have the surreal ether aesthetic. Um, and I'll start by prompting a base image using this prompt that you're going to see on the screen right now. Not going to read this for you. This is quite long and also, um, well, my English is not the best. We're going to paste this over here and we're going to click generate. A little window, it's going to pop up, which is going to be your gallery, and you're going to be able to see the results on your, uh, right-hand corner. It's going to take about a minute or two and it's going to give you now two preview options and you'll have to pick the one you like the best. But keep in mind, this is just a low-resolution preview. This is a preview. This is not the full image quality. So once you select which image you're going with, you're going to click approve.
And while that's loading, we'll start generating more images. And we need to create some variation shots, right? So, we can start with a close-up. Let's modify this exact prompt that we used, but we're just going to say, "close-up portrait of a woman." Um, going to hit generate. And then let's wait for it. This is what we get. We get a close-up of the woman. Of course, it looks similar. It doesn't look the same, but don't worry about this. Click approve. And let's generate another one. A close-up frontal image of the woman. And so we're just changing the prompt. We're just changing the keywords. So now we have a frontal image of, of the woman. So we know what, what she looks like. These variations will give, give you a tighter framing and, and more usable angles. These are all the results that we've got so far. The images do look very realistic. They do look cinematic. They have that feel to it, right? You can see. But the character, it's not identical, and that's okay. And we can fix that. However, we have to choose our best close-up portrait. The thing is, I love this one over here. I like the fact that it's closed up because we can see most of her face. I don't like the fact that her eyes are closed. So, let's just fix that using context editing.
I will go into the tools over here and I'll select context editing. I will be uploading my image that we just generated. And in the prompt section, I'm going to say, "woman has her eyes open." It's as simple as that. Um, and we're going to hit generate. Boom. Boom. And her eyes are open. Now, this looks very good. The next thing that I want to do is that I want to upscale the portrait. So, I'm going to, uh, click on portrait upscale. And then we're going to see this thing pop up. And then the next thing I'm going to do is I'm want to make sure that I'm selecting the portrait detailer mode, the fast one, and hit generate one more time. This will add texture and realism to the image. Let's wait for it to generate.
And here we have it. Her skin looks sharper. Her expression is clear. And, oh, her eyes. Just look at her eyes. Um, more, way more detailed, right? Starts to feel a bit more, more cinematic. And this looks great. Like the image looks sharp. It's good. But Cheerio, how do we animate this character in different scenes? Like, okay, we know all this stuff. Like, this is like super simple. I got you. This is, this is the sauce. Okay, so pay attention to this thing. We're going to download our detailed fix portrait. Let's call it that. And now we're going to head over to tools, video generator inside enhancer, and we're going to select image to video. We're going to upload our image and we're going to switch to the pro mode. This is very important. And we're going to select a 5-second clip or 10-second clip if we want. A 5-second clip for now does the job. And here's the magic. Here's our prompt that you do to copy and paste. You're going to say, "camera orbits around the subject. Cut to portrait of the woman looking into the camera." So now this prompt allows us to view the model's full face and give us a frame that we can screenshot and use it as a reference image to continue generation in context editing. So we're going to hit generate and we're going to see what it's going to come up with.
Okay, this is, so take a look at this. Very clean. Like we have her full face. What we're going to do is we're going to screenshot this and we're going to use it as a reference for building a series of consistent frames with the same character. And also, if you want, you can generate longer clips by adjusting the prompts. For example, I just added that the model is smiling and reaching her hand toward the camera. Right? Take a look at this. This is simple, but it gives you like a new energy to the scene. And then if you want to be using those frames to start developing your character from there, like maybe her smiling or whatever she is doing, um, you can do so. I think it's very crucial for you to kind of establish your shots in the beginning. Establish the mood and establish what is that you need for the character to do so that you can have great source images, and that is first of all, we need their full face clear visible, and maybe we need some sort of expression, right? Maybe them smiling, crying, maybe them they're reaching to, to, to something. These are great for you to determine what your source image will be. Usually, I would recommend having a full face of the model and having a model smiling, model crying, model laughing, and maybe some gestures that would be very hard to reproduce with context editing.
Now, let's get creative. And once we find our perfect frame, which is going to be this one over here, we're going to go back to enhancer, context editing, and we're going to start designing our specific scenes for our music video. This is not curated. I haven't thought about this. I'm just doing this as I go. So, obviously, the final result might not mean anything, but at least you understand what the workflow is. And this, this is the most important part right now. The rest, well, you have to show your talent there.
So, I'm going to upload my source image. Here's the image again inside context editing. And I'm going to add some prompts. Okay. So, I'm going to say, "model laying in the clouds, keeping the background unchanged." This is my prompt. Generate. Boom. Done. Easy. Look at the character. They look identical. Another one. We're going to say, "model fixing her hair." Hit generate. Boom. There's some human motion. How about we change the, the background of the scene? We can say something like, "model laying down in the grass. Dark setting, moonlight lighting the scene." Hit generate. This is phenomenal. Amazing. What if we have the model running through the forest? So, let's go. "Model running through a forest. Blue moonlight shining on her drone shot." And this one, this is going to be a longer one. I'm just going to paste it over here. I'm not going to read it all, but essentially it's going to be a portrait of a woman lying on her back, shot from a top-down perspective. Hit generate.
Here are our consistent shots of our consistent character. And we are ready to turn these into a video. Now, the very first thing that we need to do before we turn everything into a video is what? Well, you guessed it. We're going to upscale all of these images so they look more realistic. And in order for us to do that, there's two options. You can go to the upscale feature, um, on enhancer, and you can either select portrait detailer or image upscaler. The portrait detailer is going to help you detail and fix the skin texture of your subject if you have a close-up image or anything that has to do with human skin. Again, we're talking about close-up images. If the image is too far, then there's no need. We're going to be using image upscaler. The image upscaler allows you to upscale any image. It doesn't matter if it's a portrait, if it's like an object, like a, a picture in nature, it doesn't really matter. It would upscale any image. For the purpose of this video, well, we're going to be using a portrait detailer just because we're working with a real face, but both work well. We have to alternate. We have to find whatever works for our use case. Every use case is very different. We have to experiment. AI is unpredictable, right? Um, so whatever works for you works for you, and whatever works for me works for you, too, I think. Kidding.
Okay, so I'm going to upscale all these images. I'm just going to show you the before and after right here, and we're going to be ready to generate a quick video out of this. Let's generate the video.
All right, this is the part where everything comes together. I, and I cannot emphasize this enough. Your story needs continuity, a strong narrative. Right now, clips that I've shown you are, are pretty random. They are just to demonstrate what's possible. But when you actually go and make your final video, please think through it. What's your story? What do you want to communicate? What moments matter? Like ask yourself, what do I want the viewer to feel? What's happening in each frame? And why? And, and how does a scene flow into the next? Do not skip that part. It's very important because even with the best AI tools, chaotic visuals with no emotional thread would not land. Got it? Got it.
Right. So now once you've mapped out all your scenes, right, and you know exactly what you're doing, you need to write a prompt for each frame. And these prompts are your video drivers. They tell AI how to animate your image and what to focus on. And if you haven't already, I recommend you watching my other video on how to generate consistent multi-video shots. And I walk through it step by step. And now what we're going to do is that we're going to bring in all our upscale images that we have curated and created. And we're going to drop them into the video generator inside enhancer. We go to tools, video generator, and we're going to make sure that we are selecting the pro mode. Everything's enabled. We're going to drop our one of our images here and the rest of the images as well. We're going to paste in our prompts one by one for each scene. I'm going to start generating. Um, so let's just do that. I'm going to hit generate and just wait for it.
Here's the truth. Most of the work here is prompt engineering. So take your time, be intentional, work with your best friend, Chat. Um, use it to refine your language, shape your, your visual cues, use your cinematic keywords as I explained in this other YouTube video, and build out your data set, uh, shot by shot. Right? I won't spend too much time on, on how to do this. I've done it before, but here's the final result from the demo.
Oh, the stars don't shine the same. Oh, the night just calls your name. [Music] Oh, my heart's a broken shing where we fell apart.
And I think it looks great, right? Well, this took me less than 2 hours to make. Could it be better? Absolutely. Like this is, this is probably like 30% of what it could be, but I think it's strong for the time that, that I've invested into this. Now, let's compare that to an actual music video that I created. You can watch the full thing here using the exact same workflow, but with 24 hours planning, prompting, polishing, literally working on it all day. Watch Demon. See the difference that when you treat the process like art and not just automation. I hear this all the time. Hey Serio, how can I automate this brawl? You cannot automate art. You're not going to get something good out of it. You can automate it. You can automate, uh, uh, you can automate like AI avatars and bureau and stuff like that. It's just not going to be that good cuz you need this, this human input. You, you need to quality control things, right? And I believe that once you start thinking, once you start thinking like a director, like an artist, and start being intentional, the results will speak for themselves.
Now, probably you are waiting for this one. How do I put myself into the music video? How about we talk about this in the next video? Kidding. We'll just do it right now. So, in order for you to be putting yourself into a music video, we have to be using Midjourney frames. We have to train LoRA, and we have to use Enhancer and all the upscalers to finalize the process. So, let me show you a quick step by step. I'm going to keep this short and simple.
We're going to go to Midjourney, use the Omni reference feature, and we're going to upload a clear portrait of ourselves. So, in case of yourself, um, preferably with a, a plain background, we're going to set the strength to 300. Change the aspect ratio to 16 by 9. Paste in some, some cool prompts and start generating your image. This is what we get. Looks great, right? But it does not look perfect whatsoever. Um, and this is where LoRA comes in. I've trained a LoRA myself with almost 400 epochs. U, we're going to head into Flux image to image with LoRA capabilities. You can find this in Val or you can do this through ComfyUI. We're going to drop in our Midjourney image and we're going to paste in the LoRA link and set the strength and scale accordingly. I'll go with 0.4 and 1.2 and add my trigger word in the prompt section. I'm just going to hit generate.
Now this is what happens. I get an image that looks more like me. The before, the after, right? It does look a bit more like me, right? Um, and it does look a bit plasticky still. What do we do next? We'll bring this image to Enhancer and fix the skin using the new detailer mode. Or if that doesn't work for us, we can use the default skin fixer. Again, there's no one way of doing things in AI. And here are the results side by side. Once I'm happy with, with the look, um, I go back into the video editor under tools and I'm going to start prompting like I was until now. And voila, this is what you get. And that, my friend, is how you can generate a song and a music video with AI with consistent characters.
[Music] >> Your name's [Music] a broken heart. tracing the fell. >>
This was a long one and I appreciate you being here, my friend, and spending time with me. And I do hope that you learned something new today. And if you're interested in learning the best AI methods for content creation, including ComfyUI, and custom support, yes, we will do it with you and for you. I run an AI community where I help people get started and grow into high-level content producers. Here's some work from Kevin and Lucan who joined the community 3 months ago, and we host live calls, give you access to real opportunities, and you get to watch me build AI SAS projects in real time and ask any questions that you have. Thank you for being here, friend, and I want to remind you that the real magic of AI is not what it can do for you, but how it empowers you to do what you've always wanted to create without limits. This is Serio.