This AI Let’s You Make VIDEOS With FACES

We've all seen the crazy AI animations that we can make with tools like Moon Valley, AI, Pika Labs, and of course Gen 2 by Runway ML. However, there's a new game in town called Play Day, spelled Plai Day. It's made by a company called Play Labs and it can do the same type of text to video generation that we've seen from those other tools, with one major difference, which totally changes the way that people are going to start making AI videos. In fact, if I had to guess, this is going to be a lot of people's favorite video meme machine pretty soon. Now, before I really get into what this playday tool can do, let me tell you a little bit about it and why it's something you should probably be paying attention to. Play Labs, the company behind Playday, is based in LA, California and they recently received $32 million in seed funding from companies like Andreessen Horowitz, Coinbase, the talent agency UTA and Crush Ventures. And the person behind Playday, Chris DeWolf, was co founder and CEO of MySpace and Jam City. Jam City being a company behind a lot of really popular mobile games. But this feature in TechCrunch, which is actually the article that I discovered this tool from, shows us what makes this one unique over the other ones that are currently available. And it says so right here in the title. Former MySpace founders introduce a text to video generator that uses your selfie to personalize content. You can actually upload images of your face or really anybody's face and personal analyze the videos with that face embedded in the videos. It's pretty wild. Let's just go play with the tool and see what we can do with it. Now, like Pika Labs and many of the other AI generators that are out there, this one does require you to create your videos inside of Discord. Still, at the moment, it's completely free to use. There doesn't seem to be any limit on the amount of generations you can create with this tool. And I was actually surprised by how fast the generations actually come back to you. Now, looking at some of these other videos that people have created on here, we are limited to about 3 seconds with the generation, which isn't ideal. Something like Runway with Gen 2 allows us to generate four second videos, but then we can do that three more times, getting all the way up to a 16 second video. Now let's say you come across a video that you like and you're curious what it would look like with your face in it. All you gotta do is click this Personalize button and it will generate a new version. Trying to impose your face onto the video. Not always perfect. Now, one thing to note about this video is that it's actually not posting publicly to the feed. So nobody else is actually seeing this video with my face sort of superimposed into it. You can see down here in the bottom, it says only you can see this. But if I click this post, then it will actually publish this to the main feed where anybody can see it. So here's how this works. You're going to log into the Playday Discord bot. You're going to choose one of these play with AI rooms here. You've got three options to choose from. As of recording this, it's not super busy, but I do imagine once this video is live, there might be a little bit more action going on in here. The first thing we need to do is we need to train our face into the AI. So in order to do that, I'm going to type slash me. Select the me command here. And you can see it gives us the option to upload three selfies. You don't have to upload all three, but I imagine it probably gets closer to your face if you do upload all three. So I'll go ahead and select selfie one. It gives me a little box here to drag a headshot into. I have a little folder of random headshots that I've got here. So I'll go ahead and drag this one in right here. Click this little button that says plus two more, select selfie two. I'll drag this one in. Click the plus one more, select selfie three and I will drag this one in. Now there's three headshots of me for it to train on. I'll hit Enter and it's pretty dang quick. It's almost instant. Added three new photos to your personalized profile. It also says here the photos you upload won't be shown to other users only you can see this message. So nobody sees these messages that I just uploaded. Now, because I'm trained into this, I can actually go to any other video that somebody's already generated. Like this Dracula image here. Click Personalize and you can now see it says personalizing. Your video with the Playday engine will let you know when it's done. And you can see about a minute later, this started generating 8:37. This finished at 8:38. We now have Dracula with my face in it. Now, again, it's not perfect, but it definitely looks more like me than this. Original Dracula definitely added the beard and a little bit closer to my head shape. But we can also generate the videos with this that we want to swap out. And if we look at their how to use channel inside of their discord. Here's some things to know when creating videos. They give you a basic template. A medium of subject in setting, lighting, color palette, art style, details. So example, a close up portrait of a man in Mojave desert. Cinematic lighting, pastel colors, photorealistic, holding a fire torch, sunset background, High quality. Using close ups is good if you want detailed faces. For example, also suggests to improve the quality of the face. You can also start the prompt with a portrait of a man, woman, superhero, chef, etc. Use photorealistic if you want it to look more realistic, you want it to be more illustrative, use a painting or illustration of and you can try things like masterpiece and high quality to improve the quality in a general way. Let's jump back into our room two here and I will type slash lay and this is the prompt that we use to generate a video. So I will click this and now we enter our prompt. Let's just try a photorealistic close up portrait of a man driving a car. Masterpiece realistic. In my head I see the close up of the person's face and then sort of the window from the car behind them. I don't know if that's what we're gonna actually get or not, but that's what I have in mind. All right, so it took more of the side view. It did take about two minutes to generate here. It's not a straight on image, but it's not bad either. Let's go ahead and click personalize. And since my face is already trained in, it's going to take about a minute and swap out the exist existing face with my face. And one thing I have noticed and I've noticed this not just with this tool but also when using the face swapper with Mid Journey. When using things like Roop, my beard tends to cause some problems. I feel like if you're a little more clean shaven or a little bit more just stubbly or you're a woman without facial hair, it tends to look a little bit better than when I generate it with my own face for some reason. And here's what it thinks I look like driving in the car. I can see why it's not perfect because all of my headshots for me straight on and this is more of a profile shot. So hard to match that up. But like we Always say in the world of AI art, this is as bad as it's going to get. It only goes up from here. You want to get real silly with it. You can take a video of the opposite sex and swap your face on that as well. For instance, if I took this little video here, like personalize, we get this nightmare inducing monstrosity here. Now I'm trying the prompt. A photorealistic close up portrait of a man looking into the camera and sipping coffee. Masterpiece Realistic hd. Here's my video of somebody sipping coffee. Unfortunately, the coffee mug is kind of blocking the face. Let's go ahead and personalize it and see what we get out of it. And here's what that one looks like. I guess it got a little bit closer to me, but it hard to say because the bottom of the face is cut off. Now just for fun, I want to train in somebody else's face. I went and grabbed a few headshots of Elon to see if I can put Elon's face on some videos here. Let's go ahead and type me. Add selfie one, we got this image of Elon here. Selfie two, another image of Elon. And selfie three, one more. And now it's added three new photos to the profile. All right, so let's go ahead and create a new video play. I'm going to do a man in overall standing in front of a farm. Close shot. Hopefully it gets a little bit close up, but not too close up that it cuts off any of the face. All right, so here's our man in overalls in front of a farm. Although it looks more like he's flying over the farm. Let's go ahead and personalize it. See if we could get that to look like Elon. And here's Elon as the farmer. I could kind of see it just because there was a trend of people close up eating spaghetti. For a while I figured I'd do a close up of a man eating spaghetti. Here's the video I got out of that. His lips are kind of turning into the spaghetti. But let's elon ify it. And here's the Elon version of a man eating spaghetti or his face turning into spaghetti. Little more bearded than I think Elon actually is, especially at the end there. But again, this is really new technology. We've only had these text to video generators that are actually decent for maybe two, two and a half months now. And the fact that they seem to be getting better and better and now we can actually kind of Sort of swap faces into the videos. Now again, the really cool thing about Play Day is that it's totally free to use and there's no limit to the amount that you can generate. So if you really, really wanted to get a video of Elon Musk eating spaghetti, all you really need to do is generate like four or five prompts of a close up of a man eating spaghetti. Personalize all four or five of those and you know at least one or two of them are going to come out really, really solid. And that's kind of the way AI generative art has been. Most AI art tools give you four options because they know that they're not all going to be perfect, but hopefully one of the four is close to what you were looking for. So I'm going to try one last thing here and I'm actually going to put Elon's name in the prompt and then personalize it with Elon's face and see how close we get. So I'm going to this prompt a close up shot of Elon Musk drinking a milkshake. Let's add full face, see if that does anything. High quality masterpiece hd. I think it might have done like little boy Elon or something, but you can tell it's trying to make it look like Elon now. I'm curious if I actually personalize it with the Elon's faces that I uploaded. Is it going to look more like Elon or less like Elon? Let's find out. Also, another thing to note while we're waiting for this to generate, still struggles with hands a little bit. And here's the new version with Elon's face imposed into it. I definitely think it got closer. Here's the new version where I added Elon's face into the mix. Here's the original version that it generated before I added Elon's face in original version, Elonified version definitely improved when I added his face into it. But damn them fingers need work. So there you have it. That's playday, the new text of video generator on the market. Totally free to use unlimited generations at the moment, at least while it's in beta. I have had a chance to talk to some of the devs and some of the moderators inside of the Discord. They do have big plans for this. They've got some great funding, so I'm expecting big things out of this Playday app. They've already noted in some of the other articles that I read that you will be able to do things like panning and zooming and all the same stuff you can do with Pika Labs and Runway. That's all going to be available in this as well. But now we have the added benefit of customizing the faces and getting them closer and closer to who we want them to look like. And that is something that's really exciting to me. And I'm going to sound like a broken record again. This is the worst that it's going to get. We only can go up from here. Meaning that these video generations are just gonna get better and better. We're gonna be able to generate longer and longer videos. They're gonna dial in the things like the fingers and the overly close up shots and the spaghetti blending into your lips and all of that kind of stuff. Those kinks are gonna get worked out, and they're gonna get worked out sooner than anybody realizes. And I imagine in a very short amount of time, we're gonna look back and go, man, look how far this came. Remember when it couldn't generate fingers in the videos? Remember when it looked cartoony and it didn't look super realistic? Like, it's not that far off? So I'm excited to see what you go and do with this. If somebody takes this tool and puts a little bit of time into it and generates a whole bunch of prompts until they get the exact video they're looking for and then can swap it out with the exact face that they want it to look like, you're going to be able to get some real consistent characters inside of text to video generations. And that is going to be mind blowing. So head on over to Layday IO Discord and click the Join us on Discord button here. Or you can go to Discord GG Playday. It'll drop you right into this Discord where you can generate whatever you want and start tossing your face or people's faces into the videos. So have some fun with it. You like this video and you want more tutorials and to learn about new tools as they come out. And keep your finger on the pulse of all the latest AI news. Make sure you subscribe to this channel and like this video. And I will make sure more videos like this show up in your newsfeed. If you haven't Already, check out FutureTools IE. This is the hub where I curate all the coolest AI tools, all the latest AI news, and you can sign up for the completely free newsletter where I'll put those tools and all of that news directly in your email inbox. Once again, that's FutureTools IO. So thank you so much for tuning into this video. I really, really appreciate you. I'd love that thumbs up and that subscribe. It really makes me feel good about what I'm doing and that people like you appreciate these videos. I'm loving these videos and testing all these new tools and showing them off for you. Hopefully you love nerding out with me. And thank you once again. I do really appreciate you. I'll see you in the next video. Bye Bye.