初学者教程（1小时）

• Today we're discussing Weavy, an AI platform for combining various AI models under one membership.
• I'll guide you through a straightforward workflow, demonstrating its construction.
• No need to build from scratch; you can use pre-built templates or modify them as you wish.
• Understanding the structure will open endless possibilities for your projects.
• We'll explore the cost structure and features to help maximize your usage.

Today we're going to talk about Weavy. I'm going to show you a workflow. I'm going to show you the exact way I built it, and then I'm going to make it available for you if you want to go ahead and use that without having to build it from scratch. But I want to show you how it was built so you can understand how it works in case you want to modify it or create something completely different. Once you have the structure on how to build those workflows, the sky is the limit for what you can do.

So Weavy is an AI platform that combines all the AI models, all the main ones, into one single login, one single membership, one single cost. So this is like you having to go everywhere to pay your bills. You have your cell phone bill, you have your car payment, you have your credit card. Imagine having all these different logins, which we do right in here. Instead of having all different logins, all different monthly costs for AI memberships, you can go ahead and just get Weavy and everything is under one umbrella.

There's no more logins, just one monthly cost, one place, one membership, and you can combine all the different AIs under one workflow. So let's go ahead and talk about one of the most important things to consider when it comes to this platform, which is the cost. Right? So the cost is on a monthly basis. You have zero dollars to start. They give you 150 credits, which you can generate about 375 images, which is pretty good, or 25 seconds of video.

I mainly use Weavy for images and I use Google VO3 for video, and I have a different membership for that. But you can use VO3 inside of Weavy. I just choose not to, just because of cost reasons. Right. The starter plan is $24 a month, and that gives you 1,500 credits. You can do the professional plan for $45 a month, and that gives you 4,000 credits or 10,000 images. The one before is the one I have. I get around 3,700 images or 400 seconds of video, but I do my videos outside of Weavy. But you can definitely do the videos inside of Weavy.

If you don't want to create a different account and have a different membership to VO3 or Higgs Field, you can do all inside of Weavy. That's why it's awesome that you can just do that in one single login, one single monthly membership, one single platform. So now that we know all the cost information, let's dive into it and create our first workflow.

So once you are in here in your dashboard, once you log in, this is all going to be blank. All these are projects that I have already started my files. This is going to be blank for you. Up here are all the templates that they give you. So all the workflow templates already pre-built are going to be up here. And that doesn't always necessarily apply to what I'm building. So I like to know how to create them from scratch. I probably use one or two of these pre-built templates. But what I like doing is creating a new file over here and I'm going to show you the main components of a workflow. So that way you have the skill of knowing how to build a workflow and you can create infinite numbers of workflows for yourself.

So let's go ahead and create a new file here and it's going to give us a brand new canvas. Brand new canvas. And the way you manipulate this canvas on the bottom center here, you have the hand which drags it left and right. Right now you can't really see it moving because there's no element in it. So let's put an element in here so you can see it. Or as I like to call it, a node. So here's a node. Now I can move it left and right. The whole canvas is moving, not the node itself. The node itself moves through here. You grab it. Or you can move the canvas and just move it left or move it right.

To zoom out, you hit control on your keyboard and scroll on your mouse. Zoom in, Control. Scroll in, scroll out, scroll in, scroll out. So that's the basic controls of this canvas. So if I put another in here, you can see that now I'm zoomed in, I can zoom out. Hitting control and scrolling out. Hitting Ctrl on my keyboard and scrolling in. You cannot use this on the phone. It has to be on the desktop computer.

If you want to go ahead and select both of these at once, let's say to delete them, you go back here to this cursor, select that. Now you can just click on the blank spot and then select all of these. And then you can either run them. And we're going to talk about running these specific nodes in a second here. For example, let me hide my screen so you can see behind me. It says run selected. You can run it. And that's gonna run the both of them. Right now they're blank, so that's not gonna run anything. I'm going to show you how to populate the information.

So this is how you can select multiple nodes. It's just selecting the cursor here and right now if I try to move my canvas, I can't. Now if I try to move my canvas left and right like I was doing before. See, I can't do that. So the way you do it is either you go back to the hand or you just hit spacebar and hold it. And that creates the hand tool for you here. It turns it into the hand tool. So that's what I use most of the time. I just keep my left hand here on the keyboard. On the spacebar, I always have the cursor selected. And if I want to use a little hand, I just press spacebar and I can now manipulate the canvas around.

So if I'm putting multiple things in here, let's see all these different nodes. I can now move it around. I can now select everything. Move it over here, move this one over here. Individually. Zoom out, zoom in. So that's the controls. Okay, very easy. Now let's delete all of these. I'm gonna delete all of these. Hit delete on my keyboard after I selected them all. And now let's dive into the nodes on the left side panel.

These are the different AI models that you can use here. You have OpenAI, you have Gemini, you have Hicks Field, you have Sora. You have all the different image video large language models. Every AI that you can imagine are in here. So you can use that as part of your one membership cost for Weavy.

So let's dive into it. On the top, the top right, it says create. You see, Create new file. Let's click on that. And this is it. This is a brand new template. So now this is where all the tools are. It's kind of very minimalistic, but you can go so crazy with this. If you click search all the AIs that are available to you in one single scroll. So you keep scrolling, look, keep scrolling. It's in categories, right? So you have, look, you have right here, for example, video models. You have Google, VO3, Sora 2, Higgs Field, Moon Valley, Cling. And then you have here, your image models, Higgs Field, GPT Image, Imagine, Flux Pro to all the different models.

So this is what I'm saying, by you can go to one place and have your entire basket of AIs in one single platform. You don't have to go sign up for. You don't have to go sign up for mystic, you don't have to sign up for Flux, you don't have to sign up for Google, you don't have to sign up for Hicksfield. It's all in here. And also guess what, it has all the large language models. What does that mean? Like ChatGPT, Gemini.

All right, so let's start with the basics. Okay, this is it. Prompt. Do you see where it says prompt right here? I just drag and drop. This is where you type your prompt. Right now it's just a place that you can type text. For example, make me an image of a dog.

Now this is where it comes in. Where do you want to make your image at? What platform? Let's say you want to use Nano Banana. So if I can't find it here on the left, which is right here. See it's right here, Nano Banana. But if you can't find it, just search and just click nano. Just type nano and there it is. It comes up.

Okay, so this is where it all begins. This is the very basic of creating your first image inside of this platform. As you can see on the right side, I'm gonna put myself out of the way behind where I was. It says four credits. It's gonna use every time you run a Nano Banana image. Every time you create an image, four credits. And then you can have the amount of images you run here too. You know it's gonna create two images for you. That's eight credits. It's gonna keep piling up by four. So I'm gonna keep it at one.

Okay, aspect ratio. This is on the right side by the way, is where you can control your settings for this node. So every time you put a node in here, for example, let's say I want this one 2.2. See the settings on the right one 2.2. That's the settings that you can adjust for this particular AI model. So I'm going to delete that one.

So Nanobanana is right here. Now it says what's the default aspect ratio you want? I want a 9 by 16 because that's going to be my social media standing. Okay, nine by 16. So my prompt is make an image of a dog. It's kind of vague, right? Make him an image of a husky puppy at the beach.

So now I need to connect this nodes together. So I'm going to connect prompt to prompt. Let's run it. So right here it says run model. So you go ahead and click run model. So this is the Husky Puppy.

Okay guys, what do you do with this? You can do a few things. You can create different angles. This is why I love this because let me show you now what the potential of this. Okay? So now that you understand how this sort of works, I wanted to give you a quick idea of what this does. You can create something like this.

Watch this. I created a model, okay, wearing what I wanted her to wear. And then I branched out from that model and I created a close up of her. This was the one before her. This is not her. This is the one before. It's going to be this model. Okay? So once it's automated, all you have to do is it's already built. So you run. This is this model right here, right? Look how nice that came out. This is a macro shot of her. This is a close up.

Now let's get. This is very similar to this one. Let's run this one and see what this one does. Yeah. This is the model before. I was building this with, with this model and then I decided to change it. This is a medium shot. So then this is going to most likely be a shot like this.

And by the way, you don't have to do one by one. This is what you can do. Once it's built, I'm going to make this available to you and you can use it on your own without having to build it. You can use this template. So you can also do this instead of doing one by one. You see the little arrow right here on the bottom? There's a hand. The hand moves it around.

If I hit Ctrl and scroll out on my mouse. This is only available on the desktop computer. This is not available on a phone. So if I hit control on my keyboard and scroll out. Look at that. It brings it all out. If I scroll in, obviously it brings it all in. And then if I want to move it around, the little hand, what I click on to do this, I can go left, I can go right.

If you hit this arrow on the left here on the tool, I can select everything like this. Click on the blank spot. Select it like you're selecting a bunch of files. Okay. And look what happens on the bottom right. It says run selected. It says 16 credits to run selected. Why? Why does it say that? Because it's 1, 2, 3, 4, images. How many images? One, two, three, four, five. I should have been more than that. Hold on.

Yeah, 20 credits now. Okay. Selected all of them. 20 credits. Because it's five images that I selected and each one is four. Four credits. So instead of doing this, look, run, run, run. I can just select them all and run in bulk. So let me go ahead and do that for the last two here.

So I'm going to select these two. And now obviously it's going to be eight credits because it's four credits each for Nano Banana. So then Nano Banana is the image model behind this. Now when I hit run and I have 6,657 credits, you see up here on the right. So it took. It just took eight credits away from that. And there she is.

Okay, so you got a side and you got a full front profile view. And this is how I made this model, by the way. I'm going to break down all the nodes for you guys to see. Okay? This is just. Your brain is going to map it all out. Once you know the basics, you can create the same thing that I'm creating here. There are some very basic nodes that you can use.

Remember when we started here, look how basic this is. You're going to create these step by step, by step. Or you can just use my template that I made available for you. First of all, I added a prompt. Remember this? I showed you guys how to add a prompt box. You just click. You just go to the side, get the available notes, drag the prompt in there.

So I put a prompt, gender. I put another prompt, age, ethnicity, body type, clothing style. I didn't type on that prompt. This is just an instructional box. I'm not going to type the stuff that I want the results in here. I'm not going to say gender, female, male. I'm not going to type it in here. This is just a template as a guide. Think of it as a form. You're building like a form.

So this is the form fields that you want the user to input. So once you put your prompt in here, you make sure you lock it because then you don't accidentally type in there. So when you. Every node has a little thing that you can click and you can delete, duplicate, unlock, or lock, right? So yours is going to look like this lock. Boom. So I locked it. I locked the gender in, I locked the age. I locked the ethnicity.

So if I come back here a week from now and I forget that I'm not supposed to type in here, that lock reminds me, this is where you type. Now you put another prompt box. You drag it in there, the node to match it. So you got 1, 2, 3, 4, 5. So these are all prompt nodes and it says gender over here. Because you rename it.

So right here I just rename and it said prompt before and I just typed gender. And this said prompt before. I just typed age. Get it? This said prompt, I just typed ethnicity. So you rename it to match the field that you previously had inputted. That's just a renaming situation. These are all prompt notes.

Now this is where it gets a little dicey, okay? And this, I'm going to try my best to explain this to you where it's not. But once I got my fields and then once I have the actual values that I want to type in, I'm going to put all of them into a place called prompt concatenator. And the way that works is like this. I'm going to put one here. I'm going to put. I'm just going to type in prompt. It just combines everything, okay?

The concatenator combines everything. So I'm going to get the prompt concatenator when I drag it in here. And now watch what happens here. You see this? This is a brand new one. There's nothing in here. Add another input. So how many of do we have here? So you guys can see it, we got 1, 2, 3, 4, 5, 6 fields, right?

So then we have to have six fields here. We got 1, 2, 3, four, five, six. And then what we do is we're going to connect every single one of them to here. Actually, we need 12 guys because it's. Everyone has to be connected. So we're going to connect this one to one, okay? Just like it's right here. You see that, that one, that one's to one. And then we're going to connect the gender to two.

And then we're going to connect the age to one and then the age value to two to four. So 1, 2, 3, 4. And then we're going to get the ethnicity to five, the value to five. And then we need six more. 1, 2, 3, 4, 5, 6. So you guys understand, right? It has to follow this order. So you gotta get your field connected.

7. The value 8. I know it looks like a mess right now, but that's because we are doubling it up. Okay, so now let me delete that so you guys understand what's going on here. Now this is where the magic is, right? I'm going to delete this little link right here. Okay?

This is essentially what I just did. Look at this. It combines everything for you. So now the combiner connects. Now it's this text right here. I didn't add this. This is already from the values that we inserted here. So if I change this to African, watch what happens here. Look, woman, 30, let's change her age now. So all you have to do is type in your values here.

Let's make it 25 years old. So now it's 25. Essentially every node plugs into this combiner so it's combined. Okay, now what is it going to do with this information? That's when the prompt comes in. So now I just literally put a prompt in here. Prompt node, that's the prompt node that I dropped and then I wrote this.

I locked it. Let me unlock it so I can manipulate it better. Create detailed text to image prompt that will describe a model, clothing and setup for contemporary fashion campaign photo shoot. Describe the model and clothing characteristics and detail. The model should be a hyper-realistic, high-end model with futuristic. It can create your own prompt in the way the photo shoot that you want to create for whatever look you want to create. This is what I, what I have in here. This is telling essentially this combiner to do the following with this information. So then this is what we're going to do.

We're going to put the prompt in this prompt number one box right here. So now it combines it. Look, everything that we said here comes up here. And then it tells you, here are the details about the model. Perfect. This is what we want. This is essentially a prompt template with different fields that you can populate.

Okay, all right, so that is what this is. Okay, we just combined all of that. Now this is where this combiner or concatenator, this is just a dumb node. It doesn't have any intelligence whatsoever. What we need to do now is, is put in any LLM, which is essentially.

Watch, I'm going to drag this. I just typed any LLM large language model. So when I, when I click and drop it, look on the right side. Now I can use. You don't have to have a membership of ChatGPT. You don't have to have a membership of Matter AI. You don't have to have a membership of Gemini. You don't have to have a membership of Claude. It's all in here for you. So now I want this to be on ChatGPT5. So this is the brains.

Okay, so you just drop the cell. Now you pick the brains that you want on that cell. So this is what I want. I want ChatGPT5. So what happens here is I already did this up here you see, this is any NLM up here. Okay, this is what I just added here as well.

So essentially this is what I did. I just put that combined prompt into the prompt box in here. And then you hit run. I'm not gonna run it because I already have it up here. All you have to do is take the combined prompt and put it into the prompt box and then hit run. Hitting run on this is the equivalent of going to chatgpt and typing all this stuff and saying, hey, create me a hyper realistic high fashion model with this look, you know, age, whatever, this dress. And then you hit run.

So once you hit run, nothing happens here with the image yet. It's just gonna. The point of this was to create a prompt. This is just to create a prompt. Remember the instructions was to create a detailed text to image prompt. So this is what it did for us. It created a text to image prompt.

So now it says here, look, this is the text to image prompt. A hyper realistic high fashion African woman model, 25 years old, stands tall with the statuesque and powerful posture against an infinite white background. All this stuff was pre-instructed in a previous node that combined my preferences for the model. And now we have a perfect prompt to create that model for us. So then we drag this over this text node.

Okay, we drag it over to another. Now you don't. This is optional, but remember the concatenator where we talked about combining prompts? I combine that prompt. So here I'm going to bring it up again so you guys can see it. So this is how you access it. Prompt concatenator. Put it right here. And I dragged this one to. It doesn't matter, one or two, one or two.

And then up here I put another prompt box that says realistic photograph, full body. You can actually do this right here, guys, you can put that in here. So we avoid this step. But the reason why I separated was because I want to be able to customize it. If I want it to be something else other than full body, I don't want to have to go back in there and find in this mess where the full body stuff happened. I want to be able to come in here and manipulate what kind of, you know, if I want a cartoon style, I can put in here cartoon style 2D, cartoon style 3D. So I want to be able to not just create realistic photos, but also other types of styles.

Okay, so that's why I broke that down. So now that you have those two nodes connected to your combiner, right, I just marry those two. And this is just a complete dumb. There's no intelligence in this. This is just a. I call it a dumb node. It's just a combination of this and this. This is an intelligent node because it took all the instructions and when I ran it, it gave me an output.

So now we're going to take that output, combine it again with this right here, because this is where I want to be able to manipulate the style. And now we're ready to cook the image. So then I'm just going to either go to Nano Banana, so I'm going to delete this connection here just so you guys can see how I was able to get this. And again, you just go get the image model that you want. In this case, Nano Banana. Boom.

Now you just connect your prompt to the prompt of the image. So this is essentially like if you went to Gemini or nanobanana and said, hey, make me an image of this model, hit send. This is what this does. It creates the prompt for you here. And then you go ahead and put that prompt into Gemini.

Let's say you don't want Gemini to be your image model. Go on the side here, find the image model that you want. Maybe you want Higgs Field, maybe you want, maybe you want Chat GPT to be the image model. They also have an image model, remember? So then once you click on the ChatGPT node, you can connect the prompt to that.

And then when you click on it, like click in and make sure you select it. Look on the right. What kind of size do you want for your image? So we want, let's say a vertical. So that's going to be 1024 by 1536. This is essentially 1080 by 1920, right? 1080 by 1920. There's this 1024 by 1536. So let's say you want that and you want to compete, you want to see which one does better. No problem. Just put that node in there.

So here's the node for GPT and I'm going to use one for nanobanana. Okay, now let's run it. I'm going to run. This is going to cost me 8 credits. Chat GPT is going to cost me 8 credits. And Nano Banana, which is a much better one, only costs four credits. So I'm going to run that.

So this model is going to be gone after. I'm happy with this model. Then I create the different angles. You don't create the different angles before because you don't know what she looks like yet. So here. Here it is. This is Nano Banana to the rescue. Quick. And I think I asked for a dress, right? I asked for a spring dress, minimalistic. So it's going to be a minimalistic dress.

All right, let's see. This is what Chat GPT gave me. Similar, right? Chat GPT is actually better quality, man. I thought it was going to be a worse quality, but no, it gave me a better quality. So now we're going to have to make some changes here because Chat GPT won that race. And look, we can make more.

We can go ahead and say Dall E3, put it in here and then see what that gives us. Well, we're not doing that because I want to waste credits, but you get the point, right? You can put all the image models that you want to see. Mystique or mystic, however you say that, you can put all the image models that you like. Flux Pro. And then start connecting this, connect that prompt to that, connect this prompt to that.

And then let's run Flux Pro real quick. Seven, seven credits. Let's run this one to see how it looks. Maybe that's a better option for us. Let's do Flux. I'm going to delete the other one because I don't. I don't care about making more. So right now we have our model and we're going to see which image model creates a.

I forgot to change the settings. Can you change the setting on this? Yeah, I could have looked. See, this is a mistake I made. This is a one to one ratio. I didn't want that. So I have to run it again. I want to do nine by 16. I messed up. I don't want one to one ratio.

So I'm going to run it again and burn seven credits. Taking a few seconds here. Once it makes it, we're going to see which one of the three we want to branch out from. All right. Yeah, it doesn't give me the full body, but that's all right. In this case, you would want the full body.

You know why? Because it's. We don't know what. When you create the other angles here, you don't know what the bottom of the dress looks like. So you might get different bottoms of. Of different looks, different styles according to how it interpreted, you know. So even though this is a good quality one, it doesn't have the bottom and I won't feel like running it again to get the. This was supposed to give me the full body because this right here says so full body. If it didn't, it's because you have to just keep trying until you get it.

But I'm gonna leave that there so you guys can see what it did. But I'm gonna go ahead with this one right here in the middle. So I'm gonna delete. I'm going to delete this connection because this is not the we. This is the auditioning part. We auditioned the Chat GPT model. The other ones are going home.

Oh, let me see what that went. Before I do that, I forgot to. Okay. Yeah. Okay. Because once you start disconnecting, then you can get confused. So this one is going to go right here. And we're going to see what that does in a second. And then this one is going to go right here.

So I just replugged it. Okay, so this is what this is saying. It's a prompt. Okay. It's a prompt. And the prompt says, describe five distinct camera angles for this model, focusing on their face, including macro shot, close up shot, this close up shot twice. Let me unlink that.

I'm going to put medium shot here, medium shot, full body, side angle and wide shot. Perfect. All right, so we're good. So now let me lock it again. Lock it. So that's what this does. It's just a simple text field. It's not a Any LLM.

Now we're going to create the Any LLM, which is. Again, this is the smart node. It's. You can pick what large language model is going to follow the prompt that you just enter. So in my case, I want GPT5. And then I went ahead and got this prompt, connected it to the prompt box right here, and there's one called system prompt.

So if I, if I clean this up a little more so you guys can see it. Here you go, you see prompt. That's what this came from. And then the image is going to be right here because it's going to.

Now look at this image and it's going to look at this prompt and then it's going to generate five different camera angles and it's going to. This is it. This is a key. Separate each with an asterisk at the end of each prompt. This is why we're going to do that. I want to show you.

So let's run this. This has to be ran. So I'm going to run it and now it's going to look at this image and it's going to understand what I'm asking for. And it's going to give me five different angles separated by an asterisk at the end. Here it is. Macro shot. Capturing the model's eyes and smooth skin texture. Highlighting subtle details of her expression.

Boom. Separated it close up. Focusing on the model's face. Emphasizing bone structure, confidence, gaze, and natural lighting. Separated it. Very, very important that it does that. If it doesn't do that, it will do that. But if it doesn't, then the next part will get messed up.

The next part is where we're almost done. We're almost done with this. If you're following, the next part is an array. An array is on the side here. Array. So I just typed array and I dragged it. Okay, this is what an array does. Look, I'm going to do it again. So you can see it. Connect this over here. Boom.

And then look at this. Split text by. It does a comma there. See how it has all the different in. It breaks down your lines for you. So now it's breaking down all these lines, but it's not separated. Right, because it's separating it right now by a comma. I wanted to separate it by an asterisk.

Look how brilliant this is. Now watch. Look at that. Now we have 1, 2, 3, 4. Should be 5. So what happened here? Should be 5. Let's see why it doesn't do that. Okay, I see why. I see why. See, look, this one didn't finish it with the.

Unless I accidentally erased it. So I can fix that by just manually going in there and adding it. So I added it manually. I don't know if I erased it by mistake. And now we have five. Look at this. 1, 2, 3, 4, 5. So I'm going to delete that because I already have it right here.

I already have it right here. Okay. All right, so we have split text by star. We have 1, 2, 3, 4, 5. The next step, I'm just showing you, it's better if you understand how it all works. So you can create your own crazy stuff like this. You can. You can go as many ways as you want with this. This is just me showing you how I created mine.

So now we're going to do an array. We're going to do a list. Now a list. As you guys can see here, it's going to break down every single array into a single prompt box into a single node. So once I drop the list. Watch this. You drag and drop the list. 1. We need 5. 1, 2, 3, 4, 5. Right? 1. Drag and drop.

2. You get it? 1, 2, 3, 4,. In the list. Now you have a drop down in the list you have a drop down of every single one of these that is separated. You separated for you. So we have macro shot, we have close up shot. We have the medium, we have the fourth, we have the fifth.

Now essentially what this list is is a prompt box pre-filled with dropdown options. So it's a prompt box. It's not smart. It's just a complete text based prompt. Instructions. Now we're going to be able to make the images. Okay, how do we do that?

We're going to do the image model that we're going to use is Nano Banana. This is the previous model that we, we did before. So now we're going to do this new model. So the image model is Nano Banana. So I did create five of these.

So I drag and drop one, drag and drop two. You get it? So that's how I did it. I dragged and dropped five different nodes for Nano Banana. Now in here, when you click on that, look at the option 9 by 16. Look at the option 9 by 16. If it says default, what it does is it's going to default to the reference image.

Okay? If the reference image is a landscape image, it's going to give you a landscape output. So for us it's fine because the original source here is a portrait. So default is fine. This one's default. That's fine. So you know, the first, the first few ones are 9 by 16. Either way it's going to be 9 by 16 because our source image is 9 by 16.

Now this is where the magic happens. Okay, you're going to go ahead and let me erase all this stuff, guys. So you guys can see how to map this out. And the way you erase a connection, it took me forever to figure it out. You just click on it and then hit delete on your keyboard. You just click on the line, click on the line, hit delete.

Click on the line, hit delete. I couldn't figure it out for the life of me. I was trying to delete it all different ways. You have to click on the line and hit delete on your keyboard.

So now this is, this is the key right here. We're gonna select this and put it into this prompt connection of Gemini. So we hit prompt in there and we're gonna hit prompt in here. Connect that. This is so satisfying. 4, 5.

Now it doesn't know what the model looks like. It does not know what the model looks like. So this is just saying macro shot, capturing the model's eyes. It doesn't know what she looks like. So it's going to create a random person.

This is where this comes in. So this is. Was the model that we auditioned from Chat GPT. So we liked her. So now, instead of going branching out like a bunch of confusing wires to put it to plug into the connection here, where it says image, because this is the reference image.

Now this is the model that we're talking about, this girl right here. So instead of doing that and looking like a crazy spider web, we're gonna do something called a router. So a router is. It cleans it all up.

So I'm gonna drag a router over, and then with one line, I'm gonna connect that router to this image, and this is gonna split it. So it's nice and easy, nice and clean. For my timeline here, for my. For my workspace.

So the router is here. And now you can drag it as close as you want. You can put two routers, maybe, you know, if you want to put a router here and then one here so it looks cleaner, look probably better, right? Because if you just do this, it doesn't look as clean. So if you do down here, so it doesn't. It doesn't overlap with anything, and then you can route it again to that router, and then all the mess can come out of here.

So then you can plug that into here. You're gonna plug that into the second. That's it. Now, the easiest thing for you to do is understand how to control things. So for me to zoom in and out, I'm hitting control.

I'm using my wheel on my mouse, hitting control, in and out. This is how you zoom in and out. If you want to move left and right, you can go down here and click on the. On the. On the hand, and then move left and right with your mouse. Click and hold and move.

Or because I like to leave mine on the cursor here, this little cursor button, which allows me to select multiple things like this. Look, click, and drag. I can select multiple things because I don't want to have to keep going back here to the hand. All you have to do is click and hold the spacebar on your.

On your keyboard, and that turns it into the hand. So if I just either click here and move it left and right. This is the most challenging thing to get in your brain to manipulate this timeline is to really understand how to go left and right.

So if you want to keep it on the cursor and then if you want the hand, instead of going manually, you just hit the spacebar. And then look, look how it changes to a little hand. So now look, spacebar. And then I can move it around.

Now that we have everything connected, we're gonna select. Make sure this is the, the arrow. Select everything. Click on the blank spot, drag it, hold and drag it. Now look at this. 1, 2, 3, 4, 5. Every one of them is 4 credits.

4 credits. 4 credits. 4 Credits. Total cost, 20 credits. Total cost, 20 credits. Let me get out of the way so you guys can see the total cost there. Total cost right here on the bottom where I was.

So run selected. That's going to generate our model. Okay, hopefully I did this right. And it's going to generate this model right here in five different unique. There you go. It's already coming up. So now we have a close up shot.

Now this is a macro shot, right? Yeah, macro shot, close up shot, full body and then the side profile. Now you have a model. If you have a company with clothing, if you want to serve your clients, you can help them select a model that they like.

You know, select an actress to represent their brand. So now you can do this multiple angles. So by the way, I asked for a white screen. You don't have to have a white screen. I wanted to have just. It's called the infinite white background.

I wanted to have that because now you can use her as an asset. For video models like VO3. You can now upload a ingredient which is you can put a person in there so you have a consistent character. You can put an item in there so the person can be holding an item. If you want to do like user generated content for like TikTok shop and things like that, you have a spokesperson.

So here's what you would do now. So this is everything that you have. I want to redo this one actually. So this should have been a medium shot, not a full body. See, it says medium shot. Medium shot framing, shoulder to waist. Run it again.

Oh, there you go. So it was just a fluke. This is what I wanted. I'm going to show you now how are you going to use this? How are you going to use this model? So let's say you like this shot right here.

This shot right here. You like that? We need to download it on the right. Hit that right here. And then you hit download current and then you save that to your computer. If you don't want to click on the three dots and you want to export it in a different way.

You can just add another node called export. All your nodes are on the left, by the way. So export. I drag it in here and then I can connect it right here. So it's easier. Now if I want to export, all I have to do is click export. Look.

So now once you export this, which I did, this is how you're going to use it. You can either use it to post it as a photo, or you can go to Google Flow and animate it. Since I already have an image, I'm going to do frames to video I want to add. I'm going to add that model that I just created.

I'm going to change this setting to portrait. Now, we just auditioned that model. We chose the best shot that we liked of her. And guess what? If you want to get creative, guys, get a second shot of that model and do a frame, start frame and end frame. So watch this.

Let's take this frame right here. Let's take this right here. Look how nice this is. I'm going to download it. And now this is the start frame is this one. And then the end frame is that second image. Now we're going to be able to perfectly do a zoom out of her and end it right there.

So now we can say create a smooth zoom out shot for this model. No cuts. She says, My name is Zetta and this is my style. All right, let's send it. So I got first and last frame. Create a smooth zoom out shot for this model. No cuts. She says, my name is Zetta and this is my style.

While this is making this video for us, if you want to use VO3 inside of Weave, you can. Let's just delete this export. Let's say you want to make her in here so you don't have to have a VO3 membership.

V3 image to video. You see the node VO3 text to video. V3 image to video. This is what I want. Image to video. So now you're going to create the connection between her and your first frame. And then if you want the last frame, you can do the same thing I did in Google Flow.

You can take the last frame, maybe this one, and put it as your last frame. And now we need to write our prompt. So how do you write your prompt? You have to put a prompt node in here, not any LLM, because you're not asking it to have any. You're not asking it to solve any issues. You're just literally just typing.

So the prompt is what you're going to type in. So then if you go back to my Google Flow prompt, let's just copy this. Create a zoom out shot for this model. No cuts. She says, my name is Zetta and this is my style.

So we're going to copy that and then paste it right here. So this is my prompt. So I have the node and now I put that prompt into the prompt connection. So now we have the same thing without having a membership to Google Flow.

So instead of you going through here like I just did, you can just do it right inside of there. And now if you run this, it's gonna cost you 120 credits. I'm not gonna run it because I don't want to waste 120 credits.

Remember to do the image is 4 credits. And now to do this video, it's gonna be 120 credits. So you can see how that adds up. If you want to run this inside of Weavy, that's exactly how you would do that. You would connect the prompt and then your first frame and then optional your last frame.

And then you run it and then you get a video inside of Weave and then watch this. If you create, if you create multiple videos in here because you want a video for each one. Let's say I want, I want to copy this over. So I can just select it and hit Control C and then control V.

I can paste that video model in here. Look, see it? Control C, Control V. Just like you do a text. Look, it puts it right there. All right, so let's see. I want to run, let's say I want to run a video for every single one of these.

I would connect this results first frame. Results first frame. And obviously you would have a different prompt, right? You would have a different prompt for every video because it's not all going to be the same action. But let's say it is the same prompt. You would just prompt, put it into the prompt box. Or you can put that right here.

Look. So it's a little cleaner looking. Put that prompt into the prompt box, the prompt into the prompt box. So use that same prompt five times, let's say. And then let's say you want to run this. You select it all and you run.

Now you have your five videos made. It made your five videos for you. And guess what? You can, you can combine all those five videos without having to take to the video editor. Without having to take this to a video editor with five different files.

You can combine it into one file. So if you type Video. Video concatenator. Video concatenator. Right there. And then this is what, how you would do it. We need five, right? 1, 2, 3, 4, 5.

You just connect all the video outputs. Once. Once your videos make. Once they're. Once they're generated and you like it, you're going to connect all your videos to the. Just like we did the text concatenator. You're gonna do the video.

It only takes 0.1 credits to combine them. It's just nothing. So it's not even one credit. It's a tenth of a credit. So on the bottom here, it says 0.1. So once I do run model, it's essentially gonna take all these five videos and it's gonna give me one single video for that that I can then download.

So let's go ahead and see what Google Flow made us for these. Okay? And then we're going to wrap it up right here. My name is Zeta, and this is my style. Perfect. This is it.

You can either do 50 videos inside of Google Flow for $20 a month, or you can just do the videos. You don't need a membership of Google Flow if you're going to cook your videos up in here. So you can do them in here, you have to do the math to see which one is more economic for you.

You got to make that financial decision. But this is a tool that it's amazing to create photos that you can then take it to a video builder like this, like an image to video builder like Google Flow.

Okay, so going back to our model builder or character builder, let's say you want to create a piece, a short film, or just a reel for social media, and you have a specific character already in mind, whether it's a real person that you want to take a picture of or you generated the character outside of here. You don't necessarily need to build all the character information with this stuff back here.

You can simply drag and drop that character file. So I'm going to drop the file that I want to put in here. You're going to drop the file right into your workflow. See, I just dragged it from my computer and I just dropped it in here.

Now we're going to bypass this entire operation here. With this model. I'm going to delete the string that goes from here to there. I'm just going to click on the string and hit delete. And then I'm going to reconnect this guy to this string right here.

Okay. And now I'm going to delete the router string that branches out to all the image nodes and I'm going to redirect that and plug it in right over here. So now we're back to normal. So all this stuff right here, we can park it to the side.

And now all the connections are right here. So when we run this any LLM to describe five camera angles, it's going to change our array prompts. Let's run that. Because now it's drawing information from this particular character here.

So it's going to be unique to him. Now why would you want to do this again? You're going to be able to use something that you already have or a real person. You can take a picture of yourself, whether it's landscape or portrait.

Now you can have five different angles that we already created by running that. So we have five different prompts for the angles. Let's run the first one here. It's going to be macro shot, capturing fine beard and skin texture.

You see how it made that specific for this model right here that talks about his beard. So that's how you do that for a different character that you already have. So here's one. Perfect. Gave a macro shot of that.

Let's run the other one. So I'm going to click outside of the canvas, select everything and then I'm going to hit run. So now here we have our character generations. Let's see, we have a close up shot here. This one, we got an error.

So if this happens to you, just click outside of the canvas because we had everything selected before, right? So now just individually run this one again and see why that was a problem. Maybe we'll get the problem again and we won't.

Here's a side angle. This one didn't work out quite well, guys, because. Let's see, let's see why it didn't work out. I don't see legs because the image that we, we generated here didn't have legs. So maybe that's a problem.

This is wide shot place in the model in open space, face still visible, yet contextualized within minimal white background. Let's run this one again and see. See, this one got fixed. Fix this one.

Let's run this one here. I'm pretty sure because this is a wide shot and he doesn't have legs, it's gonna just make up legs for him. Maybe we'll put pants on him. Maybe it will just be his legs.

Oh, it's even smaller. So there you have it. This is. I want to make sure I leave this in here so you guys can see that sometimes there are little glitches like that. Let's just ignore this one. Let's go with something that we can work with.

Let's download this image right here. I'm going to click on the three dots. I'm going to download current. I'm going to show you what to do with this next because obviously we just.

This is just to cast the model. This is just the character casting. Okay. I'm going to download this one right here. So these two are. Okay.

What I want to do next is show you how now we can get this character and instead of dressing them up from a prompt like we did with the information back here somewhere that we said a minimal spring dress, minimalistic spring dress. Now we're going to be able to have this blank canvas here of this character.

He's wearing just a white T shirt. What if you want to put your character in a time specific wardrobe or a specific T-shirt. For example, your company T-shirt or your client's company's T-shirt that has their own logo on it. You want to be able to dress them up in an outfit that you saw online and you want to see how that would look on somebody.

So there's various reasons why you would want to customize something that he would wear from something that's real, already exists, or an inspiration that you saw on the website, for example. So once I downloaded these two images that I liked from here, I want to take you to a different workflow that I'm going to link here to the resources of this session.

And that is the multi-angle scene builder. Okay. This is the multi-angle scene builder workflow. And don't be overwhelmed by. By the way, it looks. I showed you every single thing that I used before I used here. And in addition to that, I have directions right here.

So here's the. Let's read the instructions together and I'm going to show you how this works. So enter your prompt number one. See number one. Your prompt. Enter your prompt. For example, character in the uploaded Image In a 13th century castle as a knight.

So in this case you can say whatever he's wearing. For example, here I said that he's going to be a knight, so it's going to be like a knight armor. And I also put a setting description. So it's going to be in a 13th century castle.

And because I already have a wardrobe I have in mind for him, I have an image that I'm going to upload for him to wear. I'm not going to say what he's going to be wearing. I'm simply going to say here, I want to just type this character wearing the reference image.

I'm just going to leave it like that. And you see where this is going to go. Number two, upload at least two character images as references. All right, so let's go ahead and upload the two images that we just saved.

So we're going to do this one. I literally just dragged and dropped the file from my computer. This is why you need a desktop. This does not work on the phone. And then I'm gonna get the close up shot because it will give me finer face details for this character. The other one's kind of far away.

And one more thing here it says, note the output aspect ratio of your image will be the same as the imported image. Example 69 will yield a 69 image. In this case here we have a 916 image, which means that it's a portrait image like this.

Think of it as a YouTube short or an Instagram reel. And when you're talking about 16:9, it's a landscape image, like a YouTube video, for example. So because this is a 9 by 16 image, whatever we have here is going to be the default for this output image here, the final image that we're going to get to in a second.

Because the aspect ratio on the right side where you have your settings, once you click on the node, it's going to give you settings for that node. So here I'm going to click here and here are the settings.

So by default, if you want to change it, if this image is a portrait and you want to make it a landscape, because you have in mind to make a YouTube video for it, you can certainly change the output image ratio by going to 16:9 right here, selecting it. It's right behind me.

Let me move myself here. So here's 16:9. Now when you run this, when it's ready, it's going to yield a landscape image. But let's go with 9 by 16 just in case, just so we don't have any issues. By default, images have a brain of its own.

And give me whatever image it wants, especially if you have mixed ratio images. If you have an image that's nine by 16 and then another one that's 69 and then one that's like one to one, like a square image and one that's a custom size, make sure that this instead of default has your output exactly how you want it.

So here what I want is 9 16. Now let me get back to our third step. It's actually a second step. It's part of the step number two. Upload clothing and environment or an object you want to include in your master reference image.

In this case, because I'm giving you the example of changing clothes, it's not optional for us. We're definitely going to do this. All right, so let's go ahead and upload clothing for him to wear. I'm going to get something here from my computer.

Let me see right here. I'm going to drag and drop this file right inside of that node. Here's the clothing that I want him to wear. In addition to this, guys, you can also put an environment, okay, so I left that out of this. So I can show you that you can do this part as well.

If you don't want a white background, I built this to have a white background. But if you don't want a white background, you want a certain environment. For example, anything, your backyard, a photo of a castle, whatever specific place you want this person to be in, you're going to get that image.

Let me find something that's going to be appropriate for him to be in, especially if he's wearing what he's wearing. Here's an image of an ancient place. Now we can see that this is a square image, 1024 by 1024. It's a one to one ratio.

So this is why it's important that when we go back that we have this aspect ratio exactly how we want it. Because by default, I have all different sizes. Now this is a different size, different size. It's not going to give you exactly a portrait or exactly a landscape. You have to make sure that you select that.

Now that we have that imported, we're going to see. See the node right here? It says add image output because there is no other connection that we can add. So if I drag this here, look, it doesn't have a home. You don't want it on the third one.

And there's no fourth. So you just click here and it will add a fourth one for you. Now we can drag that over here. Now, because we added another node, we have to change the prompting. This character wearing the reference image in the referenced setting, that is a setting that we just dropped.

Now that we have that, let's look at our next step, which is number three. Run this master reference image node to generate your reference image. And the reason why this is called the master reference image is because from this image we're going to generate 10 more angles.

So this is the master reference image. Okay, I can rename it right here if we want. If we rename it, I'm going to put number three, Master reference image rename. So now we have it labeled, right?

We have step one, which is right here. See the rename the rename node 1, 2, upload your character. 3. Run master reference. So that's 3 and then 4. We have another step, but we're not there yet. So let's go ahead and run this.

Now it's going to combine the character, the clothing and the background. Let's see how this is going to look. All right, this is done and it's not quite what we want. This looks silly. It looks Photoshop.

There's people in the background. This is where quality prompt comes in. So now this prompt here obviously didn't yield the results that we wanted. So we go back to the prompt and here's what we're going to do. It gave me exactly that angle of this environment and it did have people here, by the way, so I missed that.

I didn't know there were people. I don't want people to be in there. So this is what we're going to do in the prompting. This character wearing the reference image in the reference setting. I'm going to put this close, actually I'm going to do a medium shot.

I don't want the whole background to be exactly as that picture shows because it's an awkward angle. I'm going to say medium shot, he stands by the castle. And then I can say remove all people from the background.

But because I want him to be standing by that castle right here, it should be a medium shot. This is a wide shot and they placed him in an awkward position here. It looks very Photoshopped, so this does not work.

Now I'm not mentioning to remove the people because it's gonna be a close up shot. Hopefully he will be like standing right here by this castle. So let's run that again. And actually before I do that, I'm going to also change the clothing item here. I'm going to make it this very simple tunic and then I'm going to have a sash around it.

You'll see what I mean here. So now let's change the prompt here a little bit. So this character wearing the reference image in the reference setting, medium shot, he stands by the castle.

I'm going to say that is behind him. So it's not just any castle. It knows that the castle is the one from the image. And I'm going to add a sash to his tunic. And it's going to be a neutral color.

So it's not like a different color. It's going to match the tunic. Now that we have that done, let's run this node again. All right, here we have it. I had to run it a few times to get something I was happy with.

So there we have it. We have him wearing this with a sash around it. And it's the same guy. You can see it's the same character. Now all we have to do is connect the output to. This is a router.

Same thing as before. Just so I don't have a mess of webs going from here to 10 places. I just had a clean line over here. We put a router in here. See the router? Put that in there. And now the router gets routed to 10 different Nano Banana image models.

10 different ones. But before we get there, let's take a look at our instructions again. Number four. Run this large language model node to generate various image prompts.

Now we need to generate our different image prompts. What does that mean? Right here. This is the node that we're going to run. Number four. So the node has a prompt that reads this.

Describe 10 distinct camera angles. Let me zoom in so you guys can see it for the scene. Each highlighting different elements and objects from the image and capturing the look and feel. Keep every description under 30 words separate, each with an asterisk at the end of each prompt. No numbering of prefixes.

So that is the prompt that's going to go into this any LLM. Now we need to go ahead and link the image. Somehow I got unlinked. So we have to get this file into the image node so it understands that that's the image right there.

And now it's going to create 10 different angles from this image right here. The prompt only. So it's going to create 10 different image prompts for these angles. So let's run it. And that's going to separate it on the array, which will then separate it by a list. Each one will be listed in a separate prompt box here.

That's why we have the list node in here. Let's see if it already did its thing. Yep, it did a low angle shot, capturing the man's dignified stance. So now we have all different prompts here.

So we have that. Make sure it all lines up, meaning this is the fourth node. So that means it's the fourth prompt. This is the fifth node. So that means it's the fifth prompt.

It should be all lined up you line this up when you build this out. Right now it's lined up because I built it, but when I built it out, I had to manually click the dropdown because every single one of them is going to default to the first prompt.

So it's going to be the first one ten times a low angle shot right here. You have to make sure you select each one according to the order. But I already did that. So if you reuse my template, which I'll make available for you in the resources, it's automatically going to line up all the prompts in here after you run this model.

So number four is done. So let's go to number five. What is it? Run these 10 nodes to generate your output images. So I'm selecting our Nano Banana image model.

You can definitely go with another one if you want. For example, go to the menu right here. You can do flux, you can do ideogram. This is one of the best ones right here. C dream v4.

So if you want to do C Dream, you have to remap everything, meaning put 10 nodes in here, or maybe just one. You want to test it out and then you see how the mapping goes. One goes to one, two goes to two into the prompt. For example, here, let's say this one is our last one.

You put that in the prompt and then in the image node, every single box has the image that's coming out of here into this router. As this routed image output into the image so it understands what the images and what to do with the image because the prompt will guide it.

I'm going to keep this one here. I'm going to run this one to see what number 10 looks like. And I'm going to run number 10 for nano banana, which is right here. I'm not going to run all of these because it's taken up my credits and I don't need to run 10.

I just want to show you how you'll be able to use this to now generate not only a character wearing and at the same place that you described, but now you can have 10 different angles of that character wearing that in this place for your story angles.

So if you're building a story, which I'll show you in a minute, how I was able to use this in a real life situation for Instagram. So here we have it, guys. It's very dark. This angle didn't come out right. Probably because the prompting.

Let's see what the prompt said. The prompt said this number 10, a diagonal composition, sunlight splitting the frame between illuminated stone and shaded serenity. So that's what happened. It's asking for it to be like a cinematic kind of sunlight hiding him.

Let's run number nine. Just because I'm not happy with number ten. I want to show you a great example. And then here's a mistake, guys, because I didn't select the output ratio, I think I did this before.

It says default. It should be 9 by 16. This is what I wanted. Remember it said sunlight splitting. I think C Dream did a better job at splitting the sunlight, but also keeping him within the light as a shadow.

And Nano Banana gave me a very dark split. See, here it is. So this is number nine. Number nine is very similar. It's a very similar prompt and it actually came out pretty close to C Dream.

So if I ran the other ones, you can see what these were calling for. This is calling for a low angle shot. This is calling for a close-up shot, a wide shot over the shoulder shot. All different angles that you're going to get from a single image because of our prompt automation right here.

Now let me show you how to use this in the real life situation. For a piece that you want to do for your Instagram or maybe for a client. You can literally put a client's photo right here. A real person, okay?

Maybe they can be wearing this shirt already with their logo on it. That saves you the step of having to then get an image of their shirt and put it in here to then combine it. What if they were just wearing the shirt right away?

Drop their image in here and then you can place them in whatever setting you want. A construction site, candy shop, a big mansion, anything. Now this is going to combine your client with their, with their either uniform or logo or just regular clothing in that environment.

And then you can prompt by running this model 10 different ideas for shots that you will then have it for either marketing images or you can use it for video creation. Every time you create a video, it's going to be in a different angle.

So it's going to be a complete piece, a complete video. So now watch this. This is the same template, same thing that I just used, I just showed you. But this is actually something that I used and I made something for my Instagram.

I did something a little different here, right? Instead of having the different nodes for the man, what he was wearing, the setting, I made an image on Chat GPT because this image is a complete.

Just a fantasy image, right? I made this image on ChatGPT and I already had this image for a while. So I dropped it in here and I just made one clean connection from the file to my output, from the file to the any LLM that's going to look at the image and then describe 10 different angles that we can run from that image.

So essentially, I cleaned up all the inputs and I just kept the output of the image right here. It's still the same instructions. I just bypassed the first step.

Now I ran this. It gave me 10 prompts and then it gave me 10 different image angles. Gave me a wide shot. And this one's kind of weird, has some colors close up. Side angle, back angle. Look how nice this is. And then I put it all together.

I downloaded the best angles. And then I went to V03 and I prompted with the first image. Starting with the frame that I liked, I prompted a movement, I prompted actions. And this is how I was able to build this video. Right here, you can see that side angle.

I took all these images and took it to VO3 and animated them with just the prompt. And I started the frame with the image output that I was able to download from Weave. So here's the entire short piece and it has music to it. Glory to the almighty God.

So there you can see in this video got 17,000 views. 17,000. So now I see it right here on the right. 17.4. So let's go back to the workflow.

Every time I saw something I liked, I downloaded it right here. Download current frame. And once I had the frame saved, I then went to Google VO3, which we can do right inside of Weave.

Go in here and type in VO3 image to video. Because this one's text to video. We want to image to video node. Dropped it in here. Now I'm going to need a prompt.

I'm going to do is select the winner. I like this angle here. Take the results. Put it in my first frame. You don't need a last frame. You can just leave it empty.

Now let's plug in the prompt into the prompt box. And right here you can just simply say, let's see an action that makes sense here. Zoom in, shot towards the angel behind the wings. Let's write that in.

Zoom in, shot towards the angel behind the wings. So how many credits is this? This is 120 credits. So I'm not going to run this here because again, I have a membership to Google Flow and 120 credits is a lot of images for me.

So if you run this, you're going to be able to get a video that the camera zooms in past those two wings and then it closes up on the angel that's behind it. Right here. That's how our prompt works. Zoom in shot towards the angel behind the wings.

So just use common sense, plain language. Direct the camera if you wanted to zoom out. Camera zooms out to reveal angels. Wide angle shot, for example, because then it zooms out and it goes all the way back to a wide angle shot.

If you need help with prompting, I have a whole entire course on how to do video prompting for VO3 especially. So check that out if you need help with prompting. But just use plain language and direct the camera as you want it to move.

So there you have it, guys. This is the workflow for the multi-angle node. You can get a link to this template in the resources on this class. You'll be able to copy exactly what I have here, and then this workflow will live in your dashboard.

And if you mess with it, it won't change anything that I have here on my side.