9 Overlooked Google Flow Features for Better AI Videos (Full Transcript)

[00:00:02] Speaker 1: In this video, I'll show you nine powerful Google Flow features that are easy to overlook. You'll learn how to transform existing videos, create your own AI avatar, generate consistent characters and more. I'm Kevin, and let's dive in. Number one, you can transform your existing videos with AI. Here I am in a new project. I'll simply drag and drop a video from my computer into the project. And yes, I do have the rights. I'll click on, I agree. This is a video of me dunking a basketball. I know you're all very impressed. Who would have thought a tech guy could dunk? If you look closely, the hoop might not be regulation height, but let's just ignore that. Now to edit an existing video, first off, let's click on it. Then down below in the prompt field, you can describe edits that you want to make. For example, let's place this in a packed basketball arena with a crowd that erupts when I dunk. Over here, let's click on generate. And in just a few moments, Google flow creates a brand new version of my video based on that prompt. Let's check it out. That's cool. YouTuber by day, NBA player by night. Number two, and this is a quick one. You don't have to regenerate an entire clip every time you want to make a change. Here's the basketball video from earlier. Overall, it looks great, but I'd like to make one small adjustment. Right down below, let's describe the change. I want to replace the basketball with a giant chocolate chip cookie. It wouldn't be a Kevin video without that. Then over here, let's click on generate. Let's see how it turned out. It's too big for the hoop. Number three, you can create your own AI avatar. To get started in the upper right-hand corner, click on your profile picture and right here, click on create avatar. Next, let's click on get started. Here, you'll need to scan the QR code with your phone and then you can capture both your face and your voice. Once that's done, you can use your avatar in both images and videos. Let's create a video. Down below in the prompt box, click on the plus icon and right over on the left-hand side, click on avatar. Here, I see my new avatar. Click on add to prompt. Now, notice that Google flow automatically references me in the prompt. Now, right here, I could describe what I want to have happen in the scene. I'm going to teach a tutorial in the Stone Age to cavemen. Right down over here, let's now generate it. Let's now see how it turned out. Today, I'll show you how to catch a fish. Hmm. Wow. Come to think of it, I've been doing these tutorials for a long time. Number four, you can generate consistent characters. One of the biggest challenges with AI generated videos is keeping the character consistent from one scene to the next. Google flow makes this a lot easier. To get started over on the left-hand side, click on characters and right down below, you can describe a character or alternatively, you could click on upload to use a photo of an existing character. I'll create a new one right over here. I'll describe my character and then let's generate. Look at that. In Kevin Cookie Company, it includes a whisk and two cookies and cookie. That's a really creative touch. Now, in the top left-hand corner, we could give this character a name. She looks like a Sarah and right over here, I could also select a voice. Let's just go with Archanar. I think that sounds good. I'll add that to the character. Everything else here looks good. In the top right, let's click on done. Now, in the prompt field, I can simply reference Sarah just like we did with my avatar. I'll click on the plus symbol and over here, select characters and there's Sarah and that adds her to the prompt. Now, I'll type in my prompt and then let's generate it. Let's see how it turned out. Over on the right-hand side, we have the character Sarah that we generated and over on the left-hand side, we have the video of Sarah and that looks really accurate. The great thing here is it makes it really easy to reuse the same character across future scenes and that way, you keep your videos consistent. That brings us to number five. You can use start and end frames to create smoother, more controlled video generations. To get started, in the prompt box, click on the model selector and right over here, let's choose frames. To select both a start frame and an end frame, we need to switch the model to VO. Right over here, let's go with VO 3.1 quality. Omni doesn't currently support an end frame. Here, I'll click out. Now, we can upload a starting frame, an end frame or you could upload both. Now, for this example, I'll upload two different images. Here, I'll simply drag my images to the window. I have two photos. My empty chair is the starting frame. Here, I'll press and hold on it and drag it down to start frame. Over here, I have me sitting in the chair as the ending frame. I'll press and hold on that and then drag it down to the end frame. Now, all I need to do is describe how I want to get from this frame or the starting frame to the ending frame. Down below, I'll type in my prompt. I'm going to teleport into the chair in a burst of blue particles. This will be cool. Now, down over here, let's generate. Let's see how it turned out. I'm not sure why it changes my outfit halfway through, but it does end exactly on the ending frame. Number six, you can use a reference image to apply a completely different visual style to your videos. First, I'll upload the video that I want to transform. Over here, I'll click on I agree. Here, you'd see a video of me walking down a pathway. Next, I'll upload a reference image with a style that I'd like to apply to this video. In this case, I want to transform everything into a Lego world, including turning myself into a Lego figurine. To do that, we need to add ingredients to the prompt. First, I'll take this video and let's drag it down and that's our first ingredient. Next, I'll take the style and let's also drag that down as an ingredient. Now, I can type in my prompt and over here, let's try generating that. Up on top, let's see how it turned out. That's pretty accurate. Here, got the pathway color. It's a red pathway. It's got the black car in the background. It even throws in the wave like I did in the video. It has my shirt color. Although, the one thing I'm not quite sure about is why doesn't he have any hair? Hopefully, that's not foreshadowing something. Lucky number seven, you can generate multiple camera angles from a single shot. Here, I have a simple video of me working at my desk. Instead of recording the scene again from a different perspective, I can use Google Flow to generate additional camera angles. First, let's take this video and add it as an ingredient. Then, down below, let's turn on the agent and over here, I'll type in my prompt where I define the two additional camera angles that I would like. Then, over here, let's run it. Right over here, it tells me that it's going to kick off two video generations. That sounds good, so let's click on approve. Now, let's preview how they turned out. That's pretty impressive. With just one recording, I now have additional camera angles that I could use in my final edit. Number eight, you can replace the main subject in a video with something completely different. Here, I have a simple video of me trying to walk like a robot, but let's be honest, I really don't look like one. Let's keep the original motion, but replace me with an actual robot. Just like we did before, I'll take the video and let's add that as an ingredient. Then, I'll type in my prompt and over here, let's now generate. Let's kick off the generation. Let's see how it turned out. That's pretty impressive. This is the last one today. Number nine, before you generate any videos, consider using Storyboard Studio. Instead of starting with just a blank project, Storyboard Studio helps you plan your video scene by scene. To get started, over on the left-hand side, let's click into tools. Here, you'll see all sorts of different tools. Let's scroll down to the prompting category and right here, you'll see Storyboard Studio. Let's click on that. First, you can choose your storyboard style. I'll go with realistic and then let's click on get started. Here, you can either paste in a video script or at the bottom, you can simply describe the type of video you'd like to create. I want to make a documentary about baking the world's largest chocolate chip cookie and over here, let's run. Google Flow generates an initial script. Right up on top, we have the title. We also have scenes. Here, we have some dialogue and it even calls out transitions between all the different scenes. Now, I could come in here and edit any part of the script or down at the bottom, I could also use the prompt box to refine it further. For now, I think this looks good. So up on top, let's switch into assets. Here, we can see all the assets used throughout the script. We have all of the different characters. Here, we have the locations and then we also have the props. Now, you'll notice it doesn't actually show any images. That's because we haven't yet generated them. So right up on top, let's click on autofill characters to use AI to pull up an initial draft. These look great. It's now generated all of the different characters. Here, we have all of the different locations featured in the video and we also have all of the different props. If I want to make any changes, I could simply click on an asset and over here, I see the name and the visual description and I could update that and then down below, I can create a new image. But for now, I think all of these look good. So up on top, let's switch into the storyboard. Just like we did before, let's generate all of the different images. I'll click here. Google Flow now creates a visual storyboard for the entire video using all of the assets that we generated in the previous step. Since everything references the same characters, locations, and props, they stay consistent throughout the storyboard. On the left-hand side, let's click into all media. Here, you'll find all the storyboard images. From here, I can start building my video. I can right-click on any image and here, I can animate it or I could add it to a prompt to use it as reference when generating a video clip. This gives me a much more structured way to create an AI video while keeping my story, characters, locations, and props consistent from beginning to end. And that's a look at some of my favorite Google Flow tips and tricks. If you found this video helpful, please give it a thumbs up. If you want to see more tutorials like this, be sure to subscribe. Thanks for watching and I'll see you in the next video.