How to Make AI Videos With Luma Lab's Dream Machine

1. Intro: AI Videos With Luma: I'm Toma Smith. I'm an AI expert and photographer. And today, I'm talking about how to use Luma's new dream machine to create AI generated videos. It's a really cool capability, and it's something that's actually available from this new tool for free. So I'm going to walk you through exactly how to create videos, two different methods that I find that work well, how to take those videos and actually use them for things like social media. And how to extend videos and make them longer than the default five or 6 seconds that you get out of this tool. It's really pretty amazing that this technology exists, and it's out there and ready for you to try out. So let me be your guide and walk you through exactly how to make use of it. 2. How to Use Luma to Make AI Videos: Let's take a look at how to use the new dream machine from Luma Labs to create an AI video. Now, you can do this for free or you can sign up for their paid accounts, but it's super simple to actually use this and create a video. So first, we're going to go to uma labs AI. We're going to click on try now. And then we can sign in, again, either creating an account, I'd like to just sign in with G mail. I'm going to do that. Now, once we're in here, we have our main screen. And this is where we can type in our prompt for the AI video that we want to create. So there's two ways that you can approach this. One way is to prompt with just text, and the other way is to prompt with an image. So let's go ahead and say, like, I'm going to describe the shot that I want with text. So I'll say panning across a beautiful latte on a rustic wooden table at a cafe. That's a text prompt. And I want to describe here when I'm prompting this exactly what I want in my scene. So I'm going to describe the type of shot, like a pan. We're going to hopefully slowly pan across this imagined beautiful latte, exactly what I want to see, and then something about the surroundings here. You can get more complex with these. You can just keep it more general like that. If you have cinematic experience, you can describe in detail exactly the shot you want. But I find that the system does a pretty good job anyway. You can say enhanced prompt. I usually do keep that on. And try it either with or without. That's going to basically take an LLM and kind of take your basic prompt and maybe add to it and write some additional stuff and try to better interpret your intent. Again, I like to try with and without, but I'm going to start by leaving that on. Then just hit the button. And it's going to add this to the queue. Now, once it gets started on actually processing, you're going to see that change to dreaming instead of in Q. And the time that's going to take it is going to depend on whether you have a free or paid plan. I'm seeing that on the free plan, it's typically been taking a couple of hours for this to come through the queue. On the paid plan, it's usually about 5 minutes to 20 minutes depending on the complexity of the prompt. So that's again, the first approach that we can take of just prompting. You can see now, it's switched to dreaming. Prompting the text. The other approach, and I think this is great if you want more creative control is to prompt with text and image. And you can generate that image through an external AI system. So let's make a beautiful latte image with mid journey. And again, you can use N. You can actually use a real image, and then we'll upload that and add in a prompt while this is working on the first one. So here I am in mid journey in the discord. I'm going to say a matching slash, matching. I'm to say beautiful latte on a rustic. Mt. We'll let Mid Journey create some options for us here? Again, you could use Chachi PT with Dolly, you could use Diagram. You can take a real picture, whatever you want to do. I'm just going to show this with Mid Journey. Okay. So here's some options. This is a pretty good interpretation of the prompt here. I think this captures it. I'm going to go ahead and upscale choose basically this fourth image. Here it is. I'm going to download it to my computer. I'm just going to say, Save image. Save that here. It's off the screen. I'm saving it to a download folder on my computer. Now let's go back to Luma. And still dreaming about that first prompt. I'm going to go ahead and press the image button on here, and I'm going to upload that latte image that I just created. So, there we go, Latte images in, and I'm just going to say, I'm panning across this latte on the table. See very exciting. O text based one is ready, but let's do bolt at the same time. So say panning across this latte image, and then you also still can prompt with text to say what you want to have happen with that AI image that you've uploaded or any image that you've uploaded. So we're in Q. This is going to go in. You can see this took about 5 minutes to do this, and we'll come back and watch both of these in a sec. Okay, so both of these are done. I took about 3 minutes for the one generated from the AI image, about 5 minutes as we saw before for the other one. So let's have a look at these. So here's our first one, panting across the beautiful latte on the table here, a rustic wooden table. I like this shot here. I think this is indeed a beautiful latte, and I like that it's created this sort of background scene. I also like the kind of shaky camera work it's a d in here, like it does feel like a social media video. Other thing next to out on the table. I honestly don't know what that is supposed to be. It doesn't look like anything. But I think you could cut out the first couple seconds of this and use those, most likely for this video. So this is the kind of thing we're getting out of it. Again, not a perfect interpretation of the prompt intent, but still cool what it can do. Here's our other one. And you can see, it takes that latte and turns it into this big panning shot where we start with that AI generated image pretty faithfully, and then it kind of imagines, like what the rest of this room is going to look like with this little rustic table over here in the outdoors of the cafe. Pretty neat. I think you can make that into a cool vertical video, like starting with the latte pans up. And then you could have some text appear over this. Essentially, it becomes like a blank space and make a nice vertical video. Once we have these, we have two options with it. One is we can download the video. So if you have the paid version here, then there's not going to be a watermark on it. We can just say download. Now one thing I like to do is you can take these and put them into software. L CAP cut is a great one to use. So I'll go ahead and upload that video that I just downloaded. Here we are with CAP cut. I'm going to just drag this because it's not a vertical clip to make it vertical here, fill that space. Here's our video. And then I could maybe come in at this point and add some text. And let's just say Link. OMG Best latte. Social media, after all, we'll put that onto our thing here as we get up to basically having this negative space on there. Make use of that. And now we have a little video. So here's our final result there and now we get this text coming in on that video. So, as you can see, you don't have perfect control with Luma over exactly what it's going to make. Like, I decided to pan up from that cup and kind of just give us the background. But with a little editing, you can go and make that into something that's usable or interesting. This is something you could post on social. Obviously, for most platforms, you're going to want to check the box to disclose that it's AI generated. But, you know, it's a pretty beautiful looking latte on there. Pretty good for something created totally via AI. Now, the only thing you can do in Luma, we'll go back there. And we also have the option, we can press extend and describe what we want to add to this. So maybe we can say, like, cut to the outside of the cafe where this latte was served. I find this is less reliable and creating something usable, but let's go ahead and try it out. Okay, so this is done. Let's go ahead and have a look. So we start with the same video that we did before. We get there. This is the new piece. And look at that. It's going to show us the outside of that cafe, totally imagined by the software here. So I think that's really neat that it can take this AI generated image and based on the vibe and the general look. It's going to imagine what would the outside of the cafe that served that latte look like. We've got these sort of like wrought iron looking chairs, and there's some orange juice on the table. It's got this nice wooded setting, wooden roof. So I think that's a neat capability. You can extend. You can keep adding to these things. And then, as you can see, we can just keep extending this endlessly. So this is now 10 seconds long. We can add to the scene. Again, it's not always going to be perfect. You're not always going to have perfect control, but you can create longer videos by continuing to add piece by piece to your scenes. So that's a look at how to create AI videos with Luma. It's a super fun tool.