Google Flow for Beginners: Create a 2D Animated Short Film

1. Introduction: Welcome to this complete AI filmmaking Masterglass. My name is Kara tor, and I'm going to be your instruction. Right now, AI is exploding. Anyone can log into a website, type a basic text prompt and get a cool four second video clip. But I'm guessing, you have already noticed the big problem. Those clip are completely random. When people try to make a full movie, the character change faces in every scene. The lip sync looks broken, and the story falls apart. Most people are using AI like a slot machine. In this course, I'm going to teach you how to use it like a professional movie director. We are going to deep dive into Google Flow and very basics of Adobe Premiere Pro. You are going to learn the actual science behind how the AI thinks. You can stop guessing and start controlling it. I will show you my exact formula for writing prompts, how to use the AI agent tool, and most important, how to force the AI to keep your character look 100% identical in every single scene. This course is for freelancers, content creators and who want to learn AI Advanced tool like Google Gemini. So for the class project, we are not just going to talk about the theory. We are going to build something amazing together from scratch. Throughout this master class, you will be producing your own original two de animated short film called The Clockwork Sheriff. And don't worry about burning throughout all your money to learn this. I have included an entire lecture dedicated to create limited work around. I will teach you how to use zero credit images, draft smart scene economy so you can tach your digital tokens and create massive projects on a budget. Whether you are a traditional video editor looking to upgrade your skills, a storyteller who can't draw or just a complete beginner who wants to make magic, this course will give you the exact blueprint. You have the vision in your head. It is time to get into the screen. Grab a cup of coffee, open up your computer, and I will see you in the first lecture. Let's get started. 2. What is Gemini Omni: Before we start pushing buttons and making cartoons, we need to understand a little bit of magic behind the screen. Many people think AI is just a magic trick. You type words and the video pops out. But if you want to be a professional creator, you need to know how the machine actually thinks. Today we are looking at the brain behind Google Flow, Gemini Omni flash. So what is multimodal this time that Gemini and Google introduce with? The word multimodal in big letters and blow it shows Gemini Omni is something called a multimodal neutral network. That is a big fancy word. But it means something very simple. Think of how a human learns. You don't just read books, you look at pictures. You listen to sounds, and you watch things move. Gemini Omni does the exact same things. It is a single AI brain that can read, text, look at photos, listen to audios, and watch videos, all at the very same time. It doesn't just draw pictures. I understand how the real world works. It knows gravity makes things fall and it knows how light shines on a face. How does it actually create a video from sketch? It uses a process called diffusion. Let me explain it simple. Imagine a giant block of messy ugly clay. When you type a prompt like a cartoon dog, the AI looks at the block of the clay and start chipping away anything that does not look like a dog. In computer terms, it starts with pure TV static called visual noise. Step by step, it removes the noise, clearing away the mess until the exact picture you ask for appears. When you type good clear instruction, you help the AI is scout faster and better. Before Gemini Omni, Google used an older system called view. Views Great, but it was like a slow assembly line. If you wanted to change your video, you had to start all over from every beginning. Gemini Omniflas replace that old way it is lightning fast, and best of all, it is conversational. If the AI makes a character with a red hat and you want a blue hat, you don't have to start over. You just tell the AI, change the heat to blue, and it fix it instantly without changing anything else in the scene. It is like having a real world editor. Sitting right next to you. To summarize Gemini Omni, it is a smart brain that understand text, audio and video all at once. It creates art by cleaning away static until a beautiful picture is left, and it let us edit our videos just by chatting with it. Now that we know how the engine works, it is time to get behind the steering wheel. 3. Types of Video generation: Last lecture, we have learned how the Gemini omnibrane works inside. Today we are going to look at that we can ask the AI to make a cartoon for us. We will also learn how this smart AI understand the rules of the real world like gravity, space, and light. Let's start with text to video. This means we are starting from a totally blank canvas. Your tax prompt acts as your script, your camera crew, and your lightning design. If you type a sheriff walking down a dusty town street, the AI has to do, a lot of math from scratch. It has to figure out how the sheriff move, how the dust blows, and what the streets look like. The golden rule here is that you must describe the world. If you leave out a detail, the AI has to guess and sometimes it guess wrong. Now look at video to video. This is the ultimate remix. Instead of a black canvas, you give the AI of video clip that already exists to use as a guide. You can tell the AI transform into a Western Tody cartoon. Imagine you upload a real video of a person walking down a street. You can tell the AI transform into a Tody Western cartoon. The AI will keep the exact same walking movement and camera angle. But it will redo everything in that beautifully cartoon style. We use this when we want the movement to stay perfectly smooth and realistic. How the AI understand the real world. A few years ago, AI look terrible. People's arm would melt into walls and things would float in the air. Jam Omniflah vexes this beautifully. It actually understand physics and special reasoning. It knows that a heavy cot falls differently than a light feather. It knows if a character walks behind a wooden barrel, the character does not melt into the wood. They are just hiding behind it. This means our cartoon will look solid and real. Finally, let's talk about light. Lightning is a secret to make a cheap cartoon look like $1 million movie. When you tell the AI to make a scene, it builds an invisible thready words in its brain. It knows exactly where the sun is. Because of this, it knows exactly where shadows should fall on the ground. It even knows to put reflection in water or glass. So here is a protip. Always add lightning ue to your prompt, like golden or sunlight. It makes a huge difference. To summarize, we know how to start from blank Canvas that is text to video, and how to remix a clave with video to video. We also know that the AI calculates real world rules like gravity and shadow. 4. Understand Latent Space: Now that we know how the AI understand the real world, it is time to learn how it get its ideas. Today we are talking about a very cool concept called latent Ispace. What is latent Ispace? Have you ever wondered where the AI pulls it pictures from? It gets them from a place called latent space. I want you to imagine tin space as a it invisible map inside the computer's brain. This map holds millions of ideas, colors, style, and object. When you type your text prompt, you are holding a compass. Your words give the AI direction. If your directions are clear, the AI travel across the map and lands on the exact beautiful video. You are imagining. Why does the AI hallucinate? But what happens if your directions are bad? When the AI gets lots on the map, it starts to guess in AI terms. We call this a hallucination. This is when the AI generated wired glitches, melting faces or extra fingers. If you type a wage prompt like a guy in a town, the AI is confused. It thinks guy what town? Is it raining? Is it sunny? Because it is confused, the video breaks. The sect to fixing this is simple. Never let the AI guess. So the golden rule of prompt engineering is that to give the AI perfect direction, we use a trick called prompt engineering. Remember, we are not writing a cute storybook. We are writing instruction. Here are the three golden rules. First, lead with your intention. Always tell the AI the style and feeling first. Second, aim for the 20 words sweet spot. Keep your prompt short and strong. Around 20 to 30 words is perfect. Third, do not over describe. You don't need to tell the AI what a rocket looks like. I already knows. Just give it a most important details. For every single video clip we make in the class, we will use a simple three part formula. So our cartoon looks perfect. Stab one is the artistyle. We will always start by typing modern Western two D cartoon. Step two is the camera. We will tell white shot or a close up shot, and Step three is the action. Tell it exactly what our characters are doing. Like author, the Sheriff stand in the dusty street. If you put your style at the very beginning of every prompt, your whole cartoon will match perfectly. In the next lecture, we are going to open Google Flow Workspace and see where we type these prompt, see. 5. Google Flow Interface: Right now we are looking at the Google Flow homepage. As someone who spends hours editing in complex software like Premiere Pro and after effect, I absolutely love a clean, simple workspace, and Google Flow is exactly that. It doesn't overwhelm you. Let's talk a quick tour of what we see. Right in the middle, you will see a big Saxon highlighting tools from creators like you with a button to explore tools. This is a great place to check our custom workflows built by the community. Down at the very bottom, you will see your recent project galleries. If you look closely at my screen, you can actually see my already early character design. Sheet for Arthur and Pip, already sitting there, ready to go. Every time you work on something, it saves right here. You never lose it. Now let's talk about credits. Running a massive AI supercomputer takes a lot of energy. So Google uses a credit system. Think of credit system like digital token at an arcade. Every time you ask the AI to generate a brand new video, you put a token in the machine. If you look at the top, you will see a blue banner offering daily bonus of 50 karats extra. That resets every day. Google frequently runs helpful promotions like this. You will also notice a Pro badge on the top right corner next to my profile picture. Which is where you manage your account limits and keep an eye on your status. So how much does it cost generating a six second video clip? So generating a six sacking video clip cost to ten credits or eight second video clip costs you 12 adits. Four second video cost you seven carats, and ten second cost you 15 credits. All the generation depends on the generations. The carat generation also depends on the number of generation. If you choose four X, it will generate four videos simultaneously, and it will cost you 60 carats. If you generate only one output, it will cost you 15 carats. Once you click on the profile picture, you will see how many carats left in your Google account. And if you want to get carts, you can click here and you will see how many credits you will get once you pay for the carts. They have their different options or promotional options running on their website. To go to this exact website to go to Google Labs, you simply have to search for the Google Labs or Google Flow on the Google Search. That's all you have to do, and it will redirect you to this very section. Google Flow, search on Google, and you will see this. If you have never opened Google Flow or you have not created any ID using Google Flow, you will see this. This is the very first section you will see. Once you click anywhere on the Google, you will see creative Google Flow. Once you click here, it will ask you to sign in using Google account. Once you complete this step, it will redirect you to this screen. If you have a project, it will appear here. If you do not have a project, nothing will show up here, only the new project will appear. Also, if you want to see the example or the creation in the community that already been created using this very tools, you can see here. Once you click on the flow TV, you will see all these generations. Using Vo three, it is the video that is created using Video two, this video has been created. Once you click on the grid view, you will see a cluster of videos that already been generated using VO one, VO two, Vo three or Omni. If you want to toggle the switch, go to the next Nectre and you will see this. You can explore this section and get inspired. How all the videos are generated and if you want to prompt that has been used or the prompt, from which this video is created, you can simply copy the prompt from here. If you want to see the previous clip, you can watch the previous clip or if you want to watch the next clip, you can do so. And if you like this video, you can simply copy the prompt and try it yourself. If you want to create using Flow, simply create with flow. It will redirect you to new Canvas and you have to click on the new project. Once you click on the new project, this very window will appear in front of you. It really simple 6. Generate your first video: Welcome back to the studio. It is finally time to generate our first highly detailed video clip. When you open a fresh project, you will see a clean dark workspace with a small flower icon in the center. That's it, start creating or drop media. Our main focus is going to write down at the bottom of the screen. This is our text box where it asks, what do you want to create? Before we give the AI hour prompt, we need to set up our virtual camera. To do this, click the setting button located on the right of the prompt box. When the menu pops up, you will see two main tabs, image and videos. If you click the image and choose the Nano Banana two, that is the model. The image model, you will notice at the bottom that generating image cost zero credits. But today we are making a movie. Click over to the video tap. You will see option at the top four frames and ingredients. Let's set up our shot. First, select the 169 ratio. This is the option to give us a white cinematic rectangular shot. Next, make sure the AI model is set to Omni Flash. Finally, we want this clip to be a bit longer. So click on TNS for a ten second duration. Look, right below that duration button, Google Flow tells us exactly how much this cost you. Generation will cost 15 adits per video per ten second videos. Now that's our settings are logged in, click back on the main text and type or paste our highly detailed prompt. That is one shot take in the back seat of a 90s NYC taxi that has a gray care. So this is a one shot take in the back seat of 90s NYC taxi that has a gray cabra animal printed, clothes, which is quite worn. The color are muted and the light on. Through the windows, you can't see the landscape. It's probably on a city Street. There is no more lights than the night that enters dimly through the window that are slightly dirty. TA, while long hair cat and a chicken shot with a full frame camera with 85 MM lens and 2.4 focal length. The image is sharp but has a beautiful bouquet. Now let's generate it. And just like that, our video is lady. Let's click Play and see what AI has created. Look at how beautiful this is. The AI followed our instruction perfectly. You can see the worn gray Jabra print cloth on the taxi seat. Notice how the light is very muted and timid. Just like we ask, make it feel like a real night in New York City. And right there in the back seat, we have our long hair cat and our chicken. Because we ask for a 85 MM camera lens, the background is softly blurred, giving us that beautiful professional cinematic look. Now, before we move to the next lesson, I want you to point out two very important buttons at the bottom of the screen inside our text box. On the far left, there is a plus sign. If you click this, you can see you can upload your own pictures or videos. This is exactly where you would upload a video if you want to do. Video to video remix or if you want to animate the image right next to that plus sign, you will see a little button that says isn't. This is our brand new incredibly powerful feature. Think of the agent as your personal AI assistant or co writer. We just type a very long high detail prompt to get this taxi video. What if you get stuck? That is when you click the agent button instead of writing a perfect prompt, you can just tell the agent your basic ideas like, hey, help me make a video of a cat and a chicken in a taxi. Agent will jump in, chat with you, and automatically write your so let's demonstrate right here. Once you click asent and simply type, Hey, help me make a video of a cat and a chicken in taxi. Now, before it was not unable, right now, it's unable, and let's run this. So the agent will jump in, chat with you, and automatically write that complex professional prom for you. It does the hard work, so you don't have to. So it has asked me some photo realistic TD animation hand on cartoon cinematic film no. So what type of video or image Ivan. So I want is photorealistic. So whenever you feel stuck or are not sure how to describe your scene, just click the Agent button and let the AI give you helping hand. Now that we know how to generate amazing video, from scratch, it's time to build our character for our Western cartoon. I will see you in the next lecture. Agent will ask you to approve it and do not ask again. Let's approve and do not ask again feature. So it will not ask you again and again for the image or video generation. Our generation is working here, and you will see, here is the prompt that already been used. Here, this is the prompt. If you want to reuse this prompt, you can simply re use it in the prompt. And this is the agent which has created our own prompt from that single line. Let's read it. Photo realistic video of a fluffy orange cat and a white chicken sitting together in the back seat of a yellow taxicab. The cat is look out the window while the chicken looks curiously at the camera. City lights and street scene plot past the window, high quality cinematic lighting. Realistic structure and natural animal movement. Cross this one and you can see the video that Ole been generated. Both the video is generated. This is the highly detailed prompt that we have used to generate the video without agent, and this is the video that use agent, a highly detailed prompt that is created using agent detail and a simple conversational language prompt that we have given to the agent. You know that difference with the toggle off and toggle on. 7. Create your character: We know how to use Google Flow dashboard and generate a video. It's time to start our main class project. But before we can make our cartoon move, we need actors. In this lecture, we are going to learn how to design consistent character using a special image model and our new best friend, the agent. When we generated our taxi video, we use the video taps and coast it 15 carried. Character design is a process of trial and error. We want to test lots of ideas without spending our coins. This is where the image tap comes in. In your settings, you will select the model called Nano Banana two. This model is especially designed to create incredibly detailed high quality images, and the best part generating image cost exactly zero credits. We can design all day for free. Let's copy this simple one liner prompt and ask for paste, search for the Nano Banana. We do not want a video here, to generate a character, let's jump to these characters and paste it here. And let's create a character in the character sheet. So let's try with the new project. Once you click on the project, Again, the same do will pop up, and with the character, let's try this. And here you are going to create your own character. Creating a character that I can remember is tricky. We can't say a cowboy. We need a deep detail. Now let's copy this one and paste it here. And with this one, let's click on this format. Once you click here, it will add very detailed prompt within your one liner prompt. Now, read it out. Close up portrait of Arthur a two D Western cartoon. He has a very a fahy and heavy eyed and thick, dropping gray MuschePerfectly centered and forward facing. His dressed in a brown duster coat with a visible silver serif star showing his stical shoulders and chess. The style features clean two d vector art with bold line work. And flat cell shadow, illuminating by soft, even ambient light. The composition is a head and shoulder shot with clear headroom and show the character full head is entirely within the frame and not cropped by the top border. Isn't a slim less solid background. Now let's try to get using Nano Banana two. So the format will instantly turn your simple idea into a massive professional prompt that the AI understand perfectly. Let's name this character. The image has been created. Once you click on the create body, you will see the image has been already appeared here, click here and create body. With this one, let's type here a simple prompt, create a character sheet with different angle and emotions. Now let's format and again, enter the prompt here, name your character, Arthur and he sheriff. If you want to select the body, so the character sheet has been generated, and here is the portrait. Now let's select a voice. If you want to select the voice, you can select the voice. Hi there. Hi there. So I'm not selecting the image here right now. Let's make it done. With this one, our new character is ready. Once you click here, you will see body and face done with the same step. Now let's create another character, that is, we will do it again for PP. He's little robot dog. That is new character, create to deconnec robot dog. 8. Generate character with voice: We establish a rule that character needs a character sheet with multiple angles on a plain background. These prompts are engineered to force the AI to draw them from the front, side, and back. So we are going to create our second character. That is Peep, the robotic dog. This time we are going to simply copy and paste. So this time in the project, we are going to simply going to characters new character, paste this prompt and with Nano Banana two, let's create our character, and we are going to name it P P, Pep, the robotic doc. Let's try again. Want me to change. Yes, I want to change. Keep the I single. Let's change it. We can also change in our character, right? So this time, what I have typed, I type keep the I single. And if you do not like the character sheet that you want to create, then you can try it again. There are two more options you can see. Now you can see one, two, three, four, five, six. So we have a character sheet of PP. Now let's save it, and now it's done. Now, let's create a new character again. This time we are going to use CLAs. So let's try to generate CLs our main character. That is villain. Name the character here is CLEs. And if you want to select the sound for it. Hi there. Hi there. Let's add to character. Our character is generated, and it looks awesome. Let's make it done. And I have specifically asked to use this voice. Now all our character is created. Let's name this our cartoon movie. So this is our project name, cartoon movie. Okay, done. Now, once you go back to your homepage, you will see your project has name change. That is cartoon movie. And if you want to change the thumbnail of it, you can also do that. Now, with all Media and let's try to remove the agent right now with this one 9. Lip Sync Formula: So welcome back everyone. So welcome back everyone. Now that our characters are logged in, it is time to make them act. One of the biggest question I get from student is, how do I make the AI character's mouth match the words? In this lecture, I'm going to teach you the lip sync formula. First, we have to understand the problem. The AI is deaf. Video models like Google Flow generate visual based on text, not audio. If you just type Arthur says hello. The AI reads the word hello, but it doesn't actually know how long that word takes to say. A lot of times the AI will just generate a video of Arthur standing completely still. To fix this, we have to become movie director. We have to command the physical action of speaking to force the AI to animate the mouth. We use action verbs, never, say he's talking, never just say he's talking. Instead, use a strong prompt like Arthur looking angry, mouth moving actively, enunciating words clearly. Type words clearly. Always match the mouth movement to the emotion of the scene. Tell the AI, the character is shouting loudly. If it is a secret, says whispering softly, this gives the animation a much more natural realistic feel. 10. Create frame before video save credits: Welcome back to the studio. We have our character designed saved and ready to go. Now I have my Master project guide, and we are going to go scene by scene. You will get this master guide in your project material, copy our prompts and generate every single clip we need for our short film. Make sure your video settings are set to Omni Flash 16 ratio nine and 10 seconds. Let's build a movie. Let's copy this prompt and paste it here. Let me tell you to save credits, you must use image in the first generation because every time you generate the video you might get that different result. But I recommend you to use image generation first, then use this image generation as ingredients. We are going to explore it. Don't worry. So let's first generate the image of the town. Meanwhile, it copy the second sort, the hero and copy it, paste it here. This is the generation that we got. It looks good. Now, let's use this as our ingredients. Copy this prom, paste it here, at two prompt, add two prompt, and add to prompt. These three characters has been added to our prompt. Now at Arthur, this is our Arthur at two Prompt yes, it is already added to prompt, and let's make it at two prompt. This is how you can tag your character within the prompt. Now, let's generate it with video ingredients. The settings had been perfectly aligned, and now let's run it. Meanwhile, look at that beautiful establishing shot. The AI gave us those long dramatic shadows and the golden R lightning we asked for, right. It instantly set the mood. So now, the video has been generated, and you can see within the thumbnail the structure of our street, our dog, and our sheriff. Now play it along. It looks cool, right? 11. Generate first scene: Welcome back to the studio. We have our character designed saved and ready to go. Now I have my Master project guide, and we are going to go scene by scene. You will get this master guide in your project material, copy our prompts and generate every single clip we need for our short film. Make sure your video settings are set to Omni Flash 16 ratio nine and 10 seconds. Let's build a movie. Let's copy this prompt and paste it here. Let me tell you to save credits, you must use image in the first generation because every time you generate the video you might get that different result. But I recommend you to use image generation first, then use this image generation as ingredients. We are going to explore it. Don't worry. So let's first generate the image of the town. Meanwhile, it copy the second sort, the hero and copy it, paste it here. This is the generation that we got. It looks good. Now, let's use this as our ingredients. Copy this prom, paste it here, at two prompt, add two prompt, and add to prompt. These three characters has been added to our prompt. Now at Arthur, this is our Arthur at two Prompt yes, it is already added to prompt, and let's make it at two prompt. This is how you can tag your character within the prompt. Now, let's generate it with video ingredients. The settings had been perfectly aligned, and now let's run it. Meanwhile, look at that beautiful establishing shot. The AI gave us those long dramatic shadows and the golden R lightning we asked for, right. It instantly set the mood. So now, the video has been generated, and you can see within the thumbnail the structure of our street, our dog, and our sheriff. Now play it along. It looks cool, right? 12. Generate villain scene: The entry of our villain. We are going to use this scene that is the entress of our villain. Let's page the prompt with CLs at and tag this character in our character sheet. These are same now let's add CLS, add to prompt. Have you noticed how Arthur and Pep look exactly like our character that we have created, right? There is a little different with the eyes. We want a single eye, but it shows the life. And also thanks to our tags that we are using. The dust blowing in the wind make this static shot feel alive by tagging as and this small prompt that we are using, let's animate it. Because we ask for a low angle shot, the camera looks up at him, which is classic filmmaking trick to make a villain looks powerful. But at the start, you can see here is the difference. It just come into frame. But after that, you can see this is what makes it glitch. But don't worry. We can always use from this angle. So this is the perfect angle. We can use it and Bank entry of the villain. So we can remove the starting frame. Now let's generate with this one, let's try to paste here modern Western Tod cartoon, closeup of Arthur. He's our Arthur, at two prompt, he squints his eyes against the sun Ling stern and serious wind lightly blows his collar. So now let's add this two prompt. Now again, generated. Let's play this Wow. Wow. And here is the Arthur reaction. This is great close up. The AI gave him that Eastern squinting look. And the wind blowing his scholar connects the shot perfectly to the windy environment we established in scene one. So now let's move to our scene three, the standoff dialogue and lip syncing. Let's use this prompt, copy and paste. Remove this specific from the prompt and it's use the ClsA voice. Let use it. CSA voice, at two prompt. At to prompt. Now looking arrogant, laughing, while talking mouth, moving actively, nourishing words, clearly, lips moving medium close up on his face. Silsa says, This town out of time. Orman Cis says, I mistakenly tag Arthur, but we want Silas this time at to prompt. Looking arrogant. Yes, this town is out of time. Orman hand over the badge or I will take it off your ghost. Now let's generate it. This time, I don't want this one and run the prompt. 13. Correct bad video generation: So this town's out of time, old man. Hand over the badge or I'll take it off your ghost. The lips is perfectly aligned with the words. It's really awesome to see, but the problem is the output. It had been changed. So we are going to rectify this problem. To solve this problem, I simply tag this video that is at to prompt, and you can see, Hey, the video has been tagged. And now the character that we have generated. And use Silas character design sheet. And I have used a prompt, Silas character design sheet outfit on character in the video. That is, I have tagged this like this. This is what I have tagged. So this is so this is the exact poem I have used, and this is how you can actually edit your video that you have already generated. Let's play this one. This town's out of time, old man. Hand over the badge, or I'll take it off your ghost. Now, moving forward, let's use Arthur Listening. That is safety. 14. Generate our dog scene with correction: Now, copy this prompt, paste in the Omni, remove this tag. Add Arthur. Now tag Arthur, add to prompt, standing perfectly still mouth firmly close, blinking naturally listen intently medium shot. Now let's make for four second is enough with this one, and I'm using this in the prompt so that our scene always be the same. Let's run Something ain't right. So Arthur says something isn't right, right? But this is not what I wanted. But don't worry. We can use 1 second footage from the start and in the end within the middle shot. Now, let's make it turn. Let's generate another video that is modern Arthur looking stern and angry, mouth moving actively. Lips close up on his face, dramatic shadows. Arthur says, my clock still ticking and yours is about to run out. Now, let's remove this extra and change it to 10 seconds and let's. Again, for the consistency, I'm using this in the prompt. Now let's run it. So the video has been generated. Let's play this video. My clock's still ticking, Silas, and yours is about to run out. It's perfectly created the shot. Wow. The angry expression is spot on, and again, the mouth movement is continuous. We will be able to easily match the mouth movement is continuous and perfect. Now let's copy our eight shot. That is Silas saying now. Paste it here again, tag Silas, our villain. Yes, I'm going to use this one in which I'm using the audio so that it match our character. It's still standing still, mouth firmly closed, and smile, fading, listening intently, middle shot. Make it four second and run it. With this one, I'm going to use another prompt and copy paste it here in our prompt section. Modern t with PIP at the rate, I'm using the PIP at to prompt, barking aggressively small brass gear turning blue mechanical eye glowing brightly, low angle, close up, audioaut, loud mechanical barking and heavy metal gears wearing. So use this for 10 seconds and run it. So continuously with this one, let's copy this prompt shot number ten, paste it here, tag our villain. Add to prompt. And now I forgot to add this, right? With this forgotten style, you can see So how you can change the background you know already, right? What you have to do, you have to add to prompt. And this is add to prompt. Now, what you have to type, you simply have to type replace background of video with your image. That's all. So it will automatically replace the background of the video. Now, let's run it. And with this one, also, no. We need to change the background of it, right, because the background, you can see it looks modern, but the video is really amazing. So now let's use same trick here with the background to prompt. Now, let's run this again. For the character consistency and the scene consistency, you need to use the scene that you have established in the first frame. And it is really important. So now you can see the background has been replaced and I like it. So now for our peep, let's play this one. So, so. So let's try to generate it again. Let's generate this video again. And if it's not working, then I'm going to use add to prompt. And with this add to prompt, that's all. I have to do it. And I hope this time it will work. You can see the difference. It's realistic. It looks realistic, and it looks cartoonist. And I want this type of animation and graphic, not this type. The video is really amazing, but I want something different. Let's play this one. So, no, no, no, no. Now, this looks awesome. Look at Peep, the glowing eye and turning gears looks amazing in this closeup. When we take this into editing, we are going to add some heavy metal sounds effect and mechanical barking. So you can see the heavy metal sound and sound effect aligned perfectly. The mechanical baking it's really sell this movement. 15. Generate ending film scene: Now let's move to our shot number ten. We are going to use and remove this one, modern Western two D cartoon, and at Clase, that is our villain, looking surprised and scared, taking a step back hand raised in defeat, mouth slightly open in shock with white shorts, Clase says, Wow, easy. I'm leaving. Again, use this I accidentally use this prompt feature, but don't worry. I'm going to add this one and paste here again, paste Class at to prompt. Now with this one, video generator. Good. You can always change the background. I'm using a single background. You can generate another background and use that. It's your choice. Now let's again, run it. Meanwhile, it's generating we are using our 11th shot. That is our last shot. Now let's paste it here. You Class had to prompt. Away quickly down the dusty street. In the sunset, the camera is static watching him leave. With the prompt had to prompt. Again, run it. So now let's run it. Whoa, easy. I'm leaving. I'm leaving. Wow. It looks cool. Now, let's run it. This town isn't big enough for both of us. Now he walks off into the sunset. The static camera watching him Lev gives us a perfect classical Western ending of our short film. The video generation is really amazing, and I love the animation. Just like that, we have generated our entire short film. Your final task for this module is to hover over every single one of these clip and click the download button, save them all into our new folder on your computer name Clockwork Sheriff. Raw footage. So we are officially leaving Google Flow. We are going to open up Premiere Pro, import this clip, and turn them into a real movie with sound effects, music, some voice over, if needed. I will see you in the editing room. 16. Edit complete scene in premiere pro : Want to upscale it, you can upscale it if you want to download. In original, you can download in 720 P. Let's download all this video in 720 P. Now let's copy this and save every video in our computer. Download, save it, download, save I have downloaded all this video from Google Flow, and we are going to use in Premiere Pro, Clockwork Sheriff short film. Name this and copy and paste, save it in the same folder, you will find this project file in the materials section so you don't have to worry about, you can simply DC and open it with the videos that we have generated so that you can learn it easily. If you want to create your very own project, you can do so. It's totally your choice. Now let's create the project, right click New Sequence and Clockwork Sheriff. First, as per our scene, I think this is the video. Yes, right click Scale to Frame size. Also, if you want to download this Internet TP, you can also download in full HD, and it will not cost you any credit. So for this one. Yes, it looks cool with this This is the sound. Something ain't right. Yes. Done it. Move it to forward. Again, the entry of the villain. I want this. And again. This town isn't big enough for H This town's out of time, old man. You get this one? Again, this one now. I want big enough for both of us. Yes. And I want. My clock's still ticking, Silas. Or I'll take it off your ghost. The dialogue of our hero. Again, this is how you can adjust the video. Now, let's play this. This is ten second long and it looks boring, right? With this one, that's like all the videos, right click, scale to frame size, and Let's use this sound here. Increase Press for shortcut key to remove or unlink this video. Believe that you know the basic of Premiere Pro. If you know any software like the NC resolve. Final Cut Pro, you can use that also. It's totally your choice. I'm comfortable with Premiere Pro, so I'm using that. Yes, let's start with this unlink and delete this audio here, I want to use this audio. Done. Yes. Done. No? Something ain't right. Let's something ain't right. Run this here and adjust the video. This is the fim we can use. Now, let's play this. Yes. I love the entry of the villain. This town's out of time, old man. Yes. Now, again, Copy the exact sound here. This town's out of time. So at this tune, I just This town's out of time, old man. Hand over the badge, or I'll take it off your ghost. Done. Now, let's play this. My clock's still ticking, Silas, and yours is about to run out. Now, here come the bug. Done. This town. Again, done. This town isn't big enough for both of us. 49 second short film has been created. Let's play this one. Something ain't right. Wow. Wow. Wow. This town's out of time, old man. Hand over the badge, or I'll take it off your ghost. My clock's still ticking, Silas, and yours is about to run out. No. No, no, no, no, no, no, no. This town isn't big enough for both of us. 17. Project: From a blank textbook, all the way to a fully died sound design animated short film. Give yourself a massive round of applause. Think about how far we have come in this master class. You didn't just learn what buttons to press. You learn how the Gemini Omni brain actually thinks. You learn how to navigate Latent space, how to write perfect prompts, using our formula, how to command the AI to lip sync to your voice. Most importantly, you learn how to take those row AI generation into Premiere Pro and turn them into cinematic music. Now it's your turn to sit in the director's chairs. Your final project for this course is to create your very own AI animated short film. You have two choice. If you want to practice what we just did, you can use the Master Project guide. I have provided. To completely recreate the Clockwork Sheriff on your own. But if you are feeling creative, I challenge you to write your own 30 to 60 seconds story, design your own original characters using the image tab. Save them to your library, take them into your scenes, record your own voice, and ad it all the way together with sounds and effects. When you are finished, I want you to share what you made, explore your final video, and upload it to the project gallery. I want you to leave you with one final thought. AI is an incredibly powerful tool, but at the end of the day, it's just a tool. You are the storyteller. The AI doesn't have your imagination, your timing or your unique voice. It needs you to give itection. Don't be afraid to experiment. Make mistakes, use your safety nets and keep pushing the boundaries of what you can create. Thank you so much for taking this journey with me. My name is Karan, and I cannot wait to see what you create next. Keep prompting, keep editing, and I will you next time.