How to use FLORA's node-based AI canvas, from first prompt to technique

Lightly edited from the recording for readability. Wording is the speakers’ own.

Hello, thank you so much for joining. It's wonderful to kind of have you all in here. We're a large group, so I'm excited to kind of be sharing with you all, a nice cozy little introduction to the canvas. Yeah, it's, I'm super excited to kind of walk you guys through a bunch of stuff, but before anything, just a quick introduction about myself. My name is Catherine Chung. I'm a forward deployed creative here at FLORA, and what that means is I do a lot of work in regards to tool adoption for enterprise customers, as well as just talking to a lot of creatives in the community. I do a lot of educational initiatives here at FLORA, hence why we are brought here together today.

But with that being, with all that being said, before I was at FLORA, I was a brand designer at Pentagram for three and a half years, so I, it was a very intensive kind of job and experience for me, and once I kind of like learned and understood how AI was kind of coming up in the world about a year and a half to two years ago. I kind of thought that it was a good time to try and help empower other creatives in the field, making sure that no one felt kind of left behind, and so that's kind of the energy and the mindset I bring to a lot of these educational initiatives. It's to encourage as many creatives as I can in the community to, take a hold of this opportunity that doesn't come very often. So by doing that, here we are kind of looking at an introduction to the canvas with yours truly.

So with that being said, welcome to our first FLORA Lab. If we haven't met before, here's who we are. So FLORA is a node-based creative tool that brings the top creative AI models across text, image, video, audio, and 3D into one intuitive interface to accelerate creative production without compromising on craft. So FLORA has a few entry points. We have the canvas, which is a node-based infinite-style canvas that allows you to visually program repeatable workflows, freely control how media interact or generate, and allows for a more hands-on experience. That's kind of what we're going to be going through today, but two other kind of entry points to FLORA include MCP, which is using a third-party LLM like Claude or ChachiBT, one of those like desktop applications to use FLORA to execute workflows and generate images with existing contexts already built out.

And then the third is a new one. It's called create or generate a chat interface that allows you to generate images using any model we have on the platform. But today, again, we're going to be focusing on the canvas. However, stay tuned. We will be doing more of these FLORA Labs in the future. Some of them will be more about the MCP. Some of them might be about other specific topics, all that sort of stuff.

So stay tuned for things like that. So what are we going to be looking at today? Because I have you for quite a while tonight. So I'm just going to give you guys a very clear understanding of kind of what we're going to be walking through today.

So first, I'm going to be giving you a little bit of a spiel about why this is a creative advantage and why AI is not a threat. Two is going to be why nodes and preparing to use them. I have a little exercise there that'll be kind of fun. Third is going to be your first generation. A prior FLORA user.

This will be no nothing too new to you, but it'll be kind of fun to kind of just kind of revisit it as if you were taking your first generation steps. Fourth is going to be building a real workflow on the canvas. Five is going to be touring tools on the canvas. So I'll be giving you your baby steps at first, but then I'm going to show you guys, by the way, here are some other cool things that you can do on the canvas.

And then last is going to be a little bit of independent build time. And then a live canvas review. So I will be kind of going into the workspace and like peeking at people's canvases. Just keep that out in mind. If you want to work on things while you're in the workspace that pertain to work, that pertain to any personal projects, you can make your canvas private. And I will show you how to do that later.

But in the meantime, this is kind of the general agenda that will be for you. Following today. So why AI is a creative advantage and not a threat. Bear with me here.

This is all going to kind of put us all in the mindset. So I'm just going to be reading these off so we kind of get a foundational understanding of this stuff. But creative should be leading creative AI, not tech influencers on social media telling us our practice is dead because they one shot at a 30 second clip. Us as creatives know what goes behind some of the best campaigns in the world, the best websites that carry millions in traffic every day, or what goes into communicating a feeling with no words, just images.

So this is our moment. Don't let don't let Silicon Valley take this away from us creatives. Like let's let's kind of take the bull by the horns here. Another thing is designers of our caliber don't just make good looking things. We are really good problem solvers.

So I'm sure everyone in this room has been hit with this experience. But when we get hit with a last minute pivot, or a resource cut, we think of the next best solution to accomplish the task. Our creative thinking doesn't only output visual images, it outputs unique ways around problems. And that's a universal skill. We learned in design school, or on the job is now giving us an edge in adopting AI tools quicker than anyone else.

So for as long as you've been a creative, you've had to describe your idea. You've had to defend it, change it, kill it. We've all had our work torn up in front of us. Maybe you haven't good for you. I have it's fine.

But we grow that thick skin, so it's okay. So all of this has inevitably prepared you to direct an AI model to give you what you want, describe complex ideas, communicate to a client, and bring your crazy ideas to life. So with all that kind of motivational speaking, kind of just reminding us why we're here. And to get ourselves kind of inspired to start letting those creative juices flow. The next question that you might find yourself asking is why nodes compared to every other kind of AI interface. It's a natural progression of the creative process and the design thinking. We've also the following line at some point.

Sorry, my work file is so messy. When we're showing like an artboard to another creative or your art director or something. But why? Because when we are in a creative flow state, we're spitting ideas left and right coming up with new visuals faster than our hands or tools can create. Seeing our process from start to finish is how we prove our idea to the people that matter. Notes also allow for repeatable workflows and visual systems.

So making great visuals is one thing. But to systemize that approach and make it easy to repeat or pivot for a project or campaign is truly priceless. And last but not least, chat interfaces can only go so far or they can only get you so far. So some of our first interactions with AI was through a chat interface, which is great. That's awesome.

But we've moved on so quickly. We've quickly outgrown that box and we craved a more freeform way of making beautiful work without scrolling through endless iterations and finding the right version while your art director stands behind you. So that's kind of like my three main points as to why nodes can work better for some creatives compared to a chat box. Now, look, if you're a creative, you're what, Kat, I really like my chat interface. All the power to you. That's that's the beautiful thing about choice,.

But in case, since you're here exploring the idea of using nodes for your AI generative workflows, I'm happy to kind of show you around. So. I have a little mental exercise for you guys, and it's so funny because I showed this to a few of my colleagues and every single one had a very different answer. So I want to take like 30 seconds. I won't linger here too long. I want you to type in the chat. What does a blue square and a green circle equal?

I'm seeing a few things. OK, rounded rectangle. A hexagon. OK, a purple.

OK, this is a red triangle. A pink. OK, this is great. IPhone UI. SpongeBob.

This is you are all serving. Yeah, this is exactly right. Ping Pong, a nice kitty. OK, wonderful.

So, yeah,. It can literally equal anything. It can be everything that you put into the chat and more, but that same kind of equation or formula. This is kind of how it would look like in a more rudimentary kind of node based format, right?

So you would have a blue square in a node or a green circle in a node, for example, and how those two connect in these lines called noodles will kind of determine how you choose how you choose for them to interact with each other. So let's say this is my imaginary little workflow. A blue square and a green circle could simply just equal the word square in the color of the circle. It could also be the shape of the circle cut out of the blue square. Or it could be a half and half. Or it could just be a green square. The possibilities are literally endless, and this is just a long winded way of saying AI models will have a pocket full of infinite ideas or infinite possible combinations ready to whip out and throw at you at any given moment if you don't tell it almost exactly what you want.

So it's very easy to kind of prompt to the heavens about you want that green square with the slightly rounded corners, but it's giving you like an outline circle. It's where is that? Coming from? Well, it's coming from this.

So this is all to say, prompting is kind of important. And also understanding what you want out of a workflow is also very important. So let's kind of branch from here. Let's start back off with this workflow little example here where we have the blue square and the green circle equaling a green square.

So what if from here I wanted to branch off into three different directions. So maybe I have that branching off into another node, and I'm saying make this make the square spin in a dynamic way. Maybe in another node, I have changed the camera angle of the green square, or maybe in another one, it's simply as it's as simple as give the green square sharp corners. It's really nice that from here, if I needed to understand why I ended up with a spinning square, I can kind of backtrack and understand how yes, I got the green square when I combined the blue square and the green circle, and then I can kind of go from there.

So the really nice thing about this is like instead of having to kind of scroll up in a chat, copy and paste the generation into the chat again, and then re prompt it and then scroll up all the way to the top to kind of try and backtrack. It's nice that you can kind of see it all from left to right here. So now what you can do is continue the workflow continue kind of like building off from there. So you'll see here I have my green square and perspective branching off into another node. And I have it something I don't know, make the square Halloween themed. It's almost that time of year, you can see you can see where my mind is that I need it to be 50 degrees outside here in New York.

But I digress. So that's great and all. However, Catherine, this is not what I wanted. I want something more refined or graphic. It's giving spirit Halloween.

Now I have to re prompt and hope I get this right. Okay, well, I have a solution for you. It's fine. So let's say I'm okay, I'm having trouble kind of prompting my way to the perfect, Halloween themed green square. That's when I can kind of pop up a text node or use a text model to be hey, I don't know, I'm attaching this green square into this little node here, or this metaphorical node. And I'm asking it, hey, write me a prompt that would make the square Halloween themed. I want it kind of artsy, but kind of refined, but also I don't know, I don't know, Claude, what do you think?

Or I don't know, chat GPT, or Gemini, what do you think, right? And then my little LLM in that little text node will be I got you, Catherine, let me write that prompt for you. And then you can attach that text node into another image node. And that prompt will just do all the hard work for you. And you're nice, I'm pretty happy with how that came out, compared to my other spirit Halloween little square.

So this again, I'm going to show you exactly how this looks like on the canvas, but this is kind of like a, it's like a little appetizer, just kind of get the mind going. So generally, you takeaways here, before we hop into the canvas. If you can speak it clearly, and the task is straightforward slash hard to misinterpret, write your own prompt. Sometimes prompting can be as easy as we speak it. When in doubt, try to simplify the words, or try simple words. And if more context is needed, try to explain more or use a text node to help.

So as easy as I can say, make this square Halloween themed, as long as I'm okay with misinterpretation, that's fine. But if you can't find the words to explain what you want clearly, use a text node to help you write a prompt. And again, I'll show you exactly what this looks like on the canvas in a bit. So even if it takes word vomiting into a text node, if you have to dictate your wildest dreams and deepest darkest secrets to this LLM, a text node can help you get through or get a rough idea rounded into a near into a near prompt for another AI model to use.

And then if you write your own prompt at first, and you aren't getting what you want within four generations, then try using a text node to help you rewrite the prompt. My biggest piece of advice here is don't fight a bad prompt, reassess what's going wrong, and try to use a text node to write you a more specific prompt. I didn't use AI for these slides because I have a typo there. So that's how.

So, okay, Olivia, I'm going to have to ask you to resend the workspace invite back into the chat because now I'm going to show you how this all looks like on the canvas. So once you're in there, use the provided link in the chat to access the workspace for today's lab. I'm going to pause here for like a few minutes, like two to three minutes, just so everyone kind of gets themselves in there. And in the meantime, I'm going to go ahead and let's see here, what do I have on the next slide? Yeah, I'm going to wait for like about two minutes, just so people can kind of get into that workspace.

Okay. And again, if you're having any issues, try to click the link again. And if you're just joining us and you're clicking into the workspace, make sure you name your brand new canvas with your name so you don't lose it. And it's easy to kind of like search up. Pat, it looks like there's a question about best practices for naming the workspace. Do you have a recommendation? For naming the workspace for naming the canvas.

Let me see here. One second. So there's no, you can name it whatever you want. Anything that you kind of want to just like that you think is easy for you to kind of search up, but you can double click the name. I'm going to go into my private folder here. You can double click a name and then you can change the name there, or you can click the white three buttons here and then you can rename there.

Now, after you created the new project, your canvas, go into your canvas and open two image nodes. So I'm actually going to show you this right now. So I'm going to go over to an empty space. If you want to open up two image notes, here on the left corner or the left side, you're going to see this big white button with a plus sign on it. You're going to click that and you're just going to click image here.

And then you can just do the same thing. But this time, actually, if you press I on your keyboard, it will automatically populate for you. So that's a little keyboard shortcut. I on the canvas on your keyboard will pop up an image node. V will pop up a video node and T will pop up a text.

But for now, just do yourselves a favor and pop up two image nodes if you haven't already. And we're going to set both of these to Krea. So if I actually select both of these, you can change the model of both of these image nodes at once. So up here, you'll see a toolbar.

This option here where you see whatever model you can select, you can click that dropdown. And this is where you have access to all these different image models. But for now, just so everyone can like-- if you like a particular model and you're it's all right, Catherine, I know what I'm doing here, pick that model. But for those of you that just need a little bit more guidance, you can go ahead and just search up Krea to large. And you can just set both of those nodes to Krea to large.

So you should have both of these here. And so if you look at the top here, or if you look at the top right corner, they should both be set to that model here. So once we have those two image nodes-- let me go back here-- Olivia, can you paste in the prompt template that I sent you earlier today? Olivia is going to drop in a prompt template for you guys to fill out and help you run those first two image generations that you're going to get.

So for example, the template here is going to be single, think of a noun, centered low against a flat, pick a color, lots of empty space above, shot on medium format color negative film, fine grain, gentle halation, lifted blacks, and then pick a mood. This is just a template here to help give you guys a jumping off point. But let me see here. And then if you want, you could also use the example I have on the canvas here-- sorry, guys, Olivia is about to post it-- where I just have a single stack of pancakes centered low against flat, sky blue, lots of empty space. It should feel cheerful.

So I'm actually going to demonstrate what this would look like. So if I go back here-- so if I go into my image node, you'll see a little text thing will pop up. If I just go ahead and paste that prompt template, now I can go in here and be hmm, OK, a single-- I don't know. Wow. Water bottle, because I have a water bottle right next to me, picked low against a green-- a flat green background. No, not green. I don't want it to look like a green screen.

Let's say like a navy blue, lots of empty space. And then I want the mood to be horror. Yeah, let's do something like that. So let's see here.

So I'll let that generate. And then I'm also going to just-- copy and paste that into the other image node. And I'm going to change it around. So I'm going to maybe say single book, centered low against flat navy or flat red background, lots of blah, blah.

And then I'm going to write this as happy. And then go from there. Anything just once you hit-- once you're done typing in your prompt, you can just hit-- you can just hit Generate. Let's see here. Take that quick issue in the chat. It looks like they cannot copy and paste out of Zoom.

So let's just give everybody a minute to go ahead and type out this short prompt. Got it. OK. Yes, let me pop up this prompt one more time. One more time.

Let's see. There you go. I'm also going to try and make a copy of this and put it in the public workspace so you guys can kind of go in and copy and paste it. So let me try and do that. Duplicate. Paste the prompt into the Q&A under Kim's question if you're able to copy and paste it from the Q&A. It looks like it's working for many on browser.

Another thing, if you guys look in your canvas, underpinned, there should be the lab canvas that I am showing you guys. You can go into that copy of my canvas. And then you can grab the prompt from there as well. OK, I'll give that one more minute.

And then we'll move on from there. You can do-- sorry, I see LoRA's question in the Q&A. You can populate both of the image nodes with two different prompts at the same time and generate. Yeah, you can do them both at the same time. And if you're trying to look for the canvas, looks like the pins won't work there.

So let me go back here. If you search up-- what is the name of my canvas? Wow, I just forgot. Let's see. Copy of Floralab.

So if you search up Floralab projects. You should be able to find it under Floralab introduction to canvas. OK, great. I'm seeing some really cool stuff on the general workspace area.

So that's really great. If you're just joining us, Olivia, if you could post the workspace link one more time. Thank you. OK, great.

So once you join that workspace, just make yourself a new canvas, open up two image nodes, and then set them both to Krea large two, and then write in a prompt template that I provided earlier in the chat, and then go from there. But if you're just joining us a little bit late, don't worry about it. There are plenty of other things that we can work on together. But let's see here. I'm going to go back here. Make sure I'm in the one that's not set to public so I can focus on the stuff I have here.

OK. I'm seeing some really cool stuff on the canvas. Wow, I don't know. I'm seeing some really cool stuff on the workspace.

OK, great. OK, so next now we're going to take both of these. We're going to click and drag to select both of those image nodes at the same time. We're going to-- you'll see that there's a little plus sign that kind of appears when you do that.

And then you're going to extend that into an image node here. And now this is kind of that area in the kind of-- what do you call it? The little diagram I had during the presentation. So if I click and drag these, plug them into another image node. And now you can kind of get a good look at what you want out of them.

So I cheated like a cooking show. I have everything pre-baked. I kind of did this with a picture of a pancake and a picture of a glass or glasses. And here I just have these connected out into different image nodes.

So if I actually click and drag these, plug them into another image node. Oops, not a text node. Image node. This is where I can type something stack the glasses from the second image on top of the pancakes in the first image.

So you see me doing a few things here. I'm saying stack the glasses from the second image because I have my second glass-- my pair of glasses in the second image. And then I have the pancakes in the first image. So that's kind of what's going on here. If I had these kind of switched over, then I'd have to change my prompt order.

But in this case, I'm just going to leave that the same. And then I can just hit Generate here. But again, I have a few options here. One thing to know. When you guys are kind of plugging this into a new image node, you may have to set your image model back to something like Nanobanana or GPT Image. I think the best way that Krea works is without two image nodes plugged into it.

So when you're extending those original two image generations into another image node, just make sure you're on something like GPT Image or Nanobanana Pro. So I'll let that generate. But here are a few examples of those same kind of green square plus blue-- or green circle plus blue square equals either green square or like a half square half circle. These are just different ways that I kind of went about doing that, like stacking the glasses, making the glasses out of the pancake, or even just switching the background colors accordingly.

So give that a try. But in the meantime, I'm going to move on to another example. Because I'm sure a lot of us are not creating pictures of pancakes and glasses. If you are, I'm jealous and good for you. Maybe you're working with mockups. Maybe you're working with shirts or something like that.

So in this case, I have a picture of a flower. And I have a picture of a shirt. So what does a red flower plus a white shirt equal? In my case, I have-- maybe I just want the flower in front of the background that the white shirt is on.

So I just prompted that with replace the backgrounds behind the flower in the first image. That should be there. With the backgrounds behind the shirt in the second image. Another option is maybe I want the shirt to be the same red that the flower is.

So here I have made the shirt in the second image the same red color from the flower in the first image. Again, that should be there. And then here, maybe I just want the flower on the white shirt there. So while you guys are kind of generating your first two images and plugging them into another image, node to kind of create different kind of ways that they can interact with each other. Once you have a few of those, let's move on to branching off from one of those generations.

So if I revisit this little diagram I presented earlier, I have one version where the score is spinning, another version where the square is kind of in perspective, another where it's just sharp. So what I did here was I picked-- well, I like this image. What are a few other ways that I can kind of explore this? So I have that kind of connected into another image node. And if I do that, I can just kind of looks like that. And I have this one prompted with something as simple as turn this into a cinematic movie still. And it did something like this.

And I'm OK, yeah, that kind of works. Another example is plugging that into another image node and prompting change the camera angle to be from the bottom up. Which cool, kind of happy with that. And then maybe I have make the glasses float above the sack of pancakes. Give it like a little bit of a telepathy or something like that.

So pick one of the initial branches and just kind of start messing around with how to edit that existing image. So you're going from taking two images and combining them in three different ways. And then you're picking one of them. And editing it in three different ways.

So you're kind of branching off and expanding on one of those initial ideas. Another example of that in context to maybe something that we are all kind of familiar with going back to the flower and the shirt examples. These are the three ideas. These are the three initial generations that I came off of where the flowers in front of the foliage. The shirt is the red flower color.

And then the white shirt has the flower on it. So I'm OK, how about I take the red shirt here. And I want to explore three different directions. So I'm actually going to restart a few of these just to kind of demonstrate these all over again. And I'm going to slow down.

So I'm going to move these over. So let's say I'm here and I'm OK, I want to go from the red shirt and explore three different branches based off of that. Because I just explored three different ways of combining these two first images. And now I'm going to explore three different ways of branching off of it.

So I'm going to go with the red shirt. And here I kind of just focus in on this. I kind of click and drag out from one of these images. And I let go. I'm going to get this little drop down menu. And I'm going to hit image. And I think the first prompt that I wrote was pretty simple.

I think I just said make the coat hanger white, keep everything else the same. So I'm going to go ahead and generate that. I think I just have this set to Nano Banana Pro. You can mess around with different models. And I would say actually I take that back. Stay with Nano Banana Pro. It's kind of foolproof and it'll kind of save you a lot of trouble in that sense.

So I'm going to go ahead and click here and hit generate. And while that's generating with the beauty of this is I don't have to wait for it to finish. I can go ahead and explore something else. So I'm OK, well, I made that hanger white.

Now let's see if I can put this on a person. So I'm going to go ahead and do the same thing. Click and drag out from that same image node into another image node here. Then let's see what was the prompting area. In this case, I prompted put the red shirt on a full body person in the same position and environment. And so I'm going to paste that here or I can type it out, but I already wrote it.

So I think I'll just paste it and then I'll hit generate there. And so now while those two are generating, I'm what? I'm not done yet. I'm going to go ahead and drag that out into another image node like this. I'm what? Maybe I don't like the t-shirt anymore. Maybe I want this shirt to be colored shirt and kind of go from there.

So in terms of quality, in this case, I'm using nano banana pro, which has a few resolution options depending on the model. And I do have a slide that I can put up later about cats preferred models to kind of pick for this. But in general, if you're using nano banana pro for the branching off, you can set it to 4K. I would say set it to 2K for now so you don't burn through your usage in the workspace while you're here.

So try to do that. And then aspect ratio options are here as well. Right now it's set to auto. But if you wanted it landscape, or if you wanted it square, or if you wanted it vertical, you can do that as well. I'm going to just keep it at auto. You see me doing that in the top right corner. However, you can also change aspect ratios.

Here at the toolbar. So I'm going to stick to nano banana pro for now and then kind of just hit generate there. I'm going to look at this guy. What is going on here? That is not right. I don't want my guy floating in front of this foliage.

So when that happens, because it will happen, sometimes the models will just give you garbage and it's all right. It's not the end of the world. You can either. Hit regenerate, which I'm going to do here. Or I can change the model.

So even while this is generating, I can be all right, well, I'm going to give nano banana one more try. But in the meantime, maybe I'll try. See dream five pro. Or maybe I don't want to think about the model. I'm just going to go to auto select model.

This is a really great feature. If you're not sure what model to pick. So if I click this, actually, if you look on the right side here, you can pick your aspect ratio. You can pick your resolution. I think I'm just going to set this to two K and you can set your effort.

So here I'm probably going to go for a better effort here, just so I get a better kind of generation and kind of go from there. And then I'm going to hit generate one more time. And that will kind of regenerate with whatever auto picked model. That it thinks will work for my use case.

So let's see if not open at redeemed itself. I don't think so. Let's see the model. The auto select model will help.

Okay. And also, if you need to see where the model menu pops up. When you hover over a node. Here at the top of the model at the top of the toolbar. Right. If I unselect auto select model.

This is where you have selection. So all these different models. Top, like my biggest piece of advice. Do yourself a favor. Just focus on nano banana pro at 2K resolution. I'll kind of go over a few models later on as we kind of move on.

But okay. That's a little better there. That's a little better. Cool.

So that's kind of a little bit of a slowed down review of like how I branch off from one generation into another. And here are a few of these examples here as well. Now let's move on to the other version of branching off from the generations that we just reviewed. So if we look back at this diagram, what we just finished was this first big batch of generations.

So the we had the combination workflow where we're combining two assets together. And then we just finished going through branching off. Into three different directions. And then now we're going to explore branching off further from there and trying to incorporate a text node to help us write a prompt for us.

So I go back here. Yeah, so we kind of looked at these, but now let's say I have this guy here and I'm okay, I'm ready to make this. Halloween themed. I can go about that a few different ways, but I'm going to try one first where I just kind of plug that into an image node, which I'm going to demonstrate here.

And then I can type in a prompt such as. Make this Halloween themed. I want it refined and subtle through the use of greens and purples and small details that are Halloween themed. Not too, not super sophisticated, kind of straightforward.

But I got something like this, which was it's fine. I was okay, I don't think it's quite getting what I actually want. So. I'm going to demonstrate how I would use a text node to help me write a prompt.

So I'm actually going to start this fresh, but I'm going to keep this down here. So let's say. You have one of your images that you branched off from. And now you need help with a text node to write a prompt for you. I'm going to extend this out. Into a text note here. And so now what's happening is this image is serving as a reference for that text node to write a prompt for you.

I'm going to tell everyone right now, when you have this text node popped up, just set it to Claude Opus five. So you can type it or any, if you're familiar with text note or text models, you can pick Gemini, you can pick chat GPT. If you need to be told what model to pick, I tend to go for Claude Opus five. So when you have that image plugged into a text node, you can go ahead and click Claude Opus five.

So once I have that text node popped up, I can go ahead and paste the prompt in. And if I read this off to you, it's me asking the text node, write a prompt that would make the image attached Halloween themed, be creative with what effects or motives are used. Purple and green themed and a little vibrant. Again, this is me just kind of getting a gist for what I want. I can't quite specifically pinpoint exactly what I want, but I'm hoping that Claude will kind of interpret what I am going for and kind of wrap it up in a nice little usable prompt.

So once I type in what type of prompt I'm looking for, I can go ahead and hit Generate. And I'll wait for that to go. But what's going to happen once that prompt is generated is then we're going to plug this original reference image into another image node. So now I have two things branching off from my original image. I have one text node popping off from my original image.

And then I have another image node coming off of my original image. So now if I look up here again at the text node, it went ahead and wrote me this huge prompt. I'm wonderful. I would have never been able to write this myself. I'm pretty happy with that. I'm going to trust the process and go from there.

So now I'm going to go ahead and plug that text node into the image node that my original image is plugged into. So if you look here at these three nodes, I have one image on the left plugged into a text node in the middle. And then I have another image on the right that has that text node and that original image on the left plugged into it. So it should kind of look like a hill with a big box in the middle at the top.

This is kind of the workflow that you want when you're writing a prompt with another reference image to keep in mind. So a few things are going on here. Since I have that text node plugged into an image node, and that text node generated a prompt for me, you're going to see that prompt or that generated text automatically populate for you in that image node. So that's kind of what's going on here.

So I don't have to copy and paste anything. It already knows that one, the prompt was written with my original image in mind, which is my floating glasses. And now it's going to use that text node or use that prompt that it wrote for me on that image. So now I'm going to go ahead and hit generate and go from there.

So here is a pre-baked or pre-cooked version of these pancakes. So this is the same example as I just showed, but now with the final generation. And that looks a lot more successful to me. Yes, of course, it's not to each their own. I really like it serves its purpose. Would I print it on a billboard? No, but it would serve as a really great digital ad for like a Halloween themed glasses sale or something.

So pretty happy with that. But that's a really good example of, if you are finding yourself kind of limited to, this is kind of boring, this kind of drab, I need a little bit more like excitement, but I can't find the words for it. Use a text node to help you write a prompt because, oh, that looks way better. Hold on. I like that one.

Okay. Use it. You see, use text nodes to help you write a prompt. It takes a lot of the guesswork out of it.

So yeah, I love that. I'm going to mark that. By the way, a little other like side note here. If you really like a generation and you want to make it really easy to find in a big canvas, up at the toolbar, when you hover over a node, you'll see a few options here, but pay attention to this little circle here where it says tag. If I click that, you can actually tag your image nodes or whatever to whatever color you want.

So since I like this so much, I'm going to mark it green. And now if I zoom out, since this canvas is so big and it has so many image nodes on it, if I go to my title, the title of the project up here at the top left corner, you're going to see a green circle there. And if I click that, oh no, I could, oh yes. I could see that green marked node highlight on my canvas.

So it's easy to kind of look at. So if you find yourself like trying to look around for the nodes that you're trying to look for, marking them or tagging them makes it really nice to kind of just go back and navigate. Great. So let's see here.

Let's see. We're one hour in. I think I have, let's see here. Okay, great.

So let me show you an example that might be a little bit more realistic. Let's stick to you guys. Again, going back to my shirt, man, these came out so funny. You're going to have a laugh.

Okay. Here's an example where I took my guy. I was okay, I love this. I love this shirt on this guy, but I also love my button up shirt.

So let me see what this guy would look like with the button up shirt. So this is me repeating. If I go back here, remember when we plugged in the square and the circle? I'm doing the same thing with two of the generations that were branched off from one of those other, from one of those other branches.

So again, here, if I look at these two images, I have a man with a plain t-shirt, and then I have a collared shirt here. I just have both of those combined into an image node. So I'm actually going to click and drag this out. Into an image node. And again, as simply as I can voice it, I don't need a text note for this. It's just simply changed the shirt from the second image to the shirt in the first image.

And then I can kind of go from there. So now I have my guy in that collared shirt. And here's an example where I'm using a text node. So just like before, I plugged that into a text node and I said, let's see here, write a prompt that will create a single cinematic high fashion shot of the image attached. The camera angle should be dynamic and editorial. Pose can be more active, et cetera.

So here, instead of prompting something make this Halloween themed or make this yada, I'm taking a little bit more of a realistic approach, such as, okay, let's kind of make this as a cinematic high fashion shot of an image. So I kind of wanted to see what cloud would write me because I would never be able to write, or I could, if I spent like 30 minutes sitting down and being okay, pose, he's captured mid motion body turned three quarter away from a source. So it's fine. Cloud wrote that for me. I could never.

So this is an example again, where I'm plugging an image into a text node to write me a prompt. And then again, I have that text node in an image node with my original reference plugged in. Look at these are so, this man is mysterious. He's you like my shirt? Not quite my tempo, honestly.

So it was really funny. Like in this case, I don't know what I was exactly kind of like looking for. I kind of, I got a really like static picture of a man. And then I asked it, make it editorial, make it Vogue. My awkward man in an editorial pose. And what? That's my fault.

I can't even blame the AI. It's fine. So it's one of those funny situations where all of a sudden you're damn, maybe this idea, isn't looking right. And what? It's okay, because now I can kind of revisit how to go from here. Maybe I replaced the man with a different looking man. Maybe I changed the shirt and rerun it, all that sort of stuff.

So I love to show the both really good generations that I do. And also the bad ones, because we all kind of encounter that no matter how long we've been doing this. So it's the same thing. Like after three and a half years at an agency, if you asked, if you like sat me down in front of illustrator and like asked me to design a star logo, I would be tripping. I'd be I don't know where the pen tool is all of a sudden.

So still happens. It's fine. We're all human. But I have another example here.

This is just kind of the same sense, just with two different images. And I won't regenerate any of these. It's just kind of a rinse and repeat of everything I just showed you, such as, combining two images like this, Arizona image and this image of two shoppers together. Maybe you're doing something in Arizona. I don't know, but I have these two images plugged into an image note up here.

And then I'm okay, apply the first image to the background white space of the second image. Because if you look at this image here, there's a white space there. So that's kind of what's going on there. And then another one was where I plugged those two in an image note. And I said, take the shoppers from the second image and placed them walking in the path in the first image.

And then another image where I created an art installation. I think I prompted create an art installation in the middle of the desert from image one. That displays the shopping culture of America use the second image as inspiration for the art installation. So, a few things that I kind of branched off from here was I took this middle image here, plug those into a few image notes.

And then I said one, we're okay, create a closeup of those two people. So I can kind of focus in on those two a little bit more here. I was what changed the people's outfits to be more Arizona desert themed. And so now she's wearing a genius skirt and like cowboy boots.

And then this one I simply just asked, changed time of day to golden hour and got something like that. So lots of different ways. You can kind of plug these in with each other here. I took the image on the left and then I changed their outfits on the right to the Western outfits from this image.

So as you kind of see me like plugging things in, you see the noodles kind of crossing in a lot of different directions. Feel free to kind of do that same thing. The combination of two notes that you did in the beginning is not exclusive to the beginning of a workflow that can be carried into any part of a workflow in general. So that's kind of how I'm trying to like teach you guys, just getting a sense of how things can connect and how things can kind of influence each other, in a very kind of like base level way.

So that connecting things together and influencing things with each other, kind of becomes the basis of your understanding of how these nodes interact, more than anything because picking a model, honestly, it's easy. These models are getting so good that honestly, you can one shot a lot of these images just by writing the right prompt with a really strong, image model, but the concept of kind of taking two things and having them influence each other is a little bit more of the underlying value of kind of doing things in a node based canvas. Okay. So a common question coming through real quick. Could you please repeat how you made four generations at the same time and how you can adjust the number of generations? Yes. Thank you for that.

Perfect. So let's say I'm not going to give this man another try. I'm not, I'm going to revisit, I'm going to revisit my Halloween glasses. Okay.

So I'm going to actually isolate this out so I can focus here. Okay. So let's say I have my image on my left and then my text node here, just going to plug this into another image node with my text note plugged in. So I'm here and I'm okay, great. My prompt is plugged into this image node. I'm ready to kind of generate what this might look like. When I hover over this arrow, you're going to see a few options here or actually just one option.

You're going to see a number, a minus sign and a plus sign. When I click the plus sign, this is determining how many versions or how many interpretations of this prompt do you want to see? So in this case, I could do nine. I could do 10. I could do up to 20. Maybe don't do 20.

So you don't burn through your usage in the workspace, but I'm going to do four. So if I do that, I can hit generate. And then I'll see four, nodes pop up on the canvas. Now, if you do not see four on your canvas, I'm actually going to move these over here. I'm going to copy this right here. Sorry. One second.

If you do not see four nodes pop up on your canvas, that's okay because it's probably popping up like this, within your note history here. So if you see that it's the same thing, it's almost as if you clicked the generate button four times. That's what the variation button will do. There is a preference in your settings, which I can show. I think if I click my profile image here and then I go to my preferences and then you'll see in your preferences, there's variations and then variation behavior. Right now I have mine set as on canvas so that I see my nodes. Populate on canvas, when I select like that variation, but you can also do in node.

If you don't want things to kind of fly onto your canvas at will. So if you have a preference, you can change out in your preferences. But what that does is just gives you, some of these are very similar. Like here, the pumpkin is kind of actually, this is kind of why I don't like the in no generation. Cause I can't see them side by side and I'm a very visual person.

But if I go back here, now I can see. Okay, these glasses are at an angle. These glasses are just still, these glasses have pumpkin or jack-o'-lantern faces in them. All that sort of stuff. And it's the same prompt, just different interpretations.

So let me go just move those. Okay. I am going to pause here and now I know Olivia and DJ are working overtime picking up on these questions. So I'm just going to hop in here and just take a look at things.

Let's see here. Okay. And then we will have a 10 minute break, just to kind of absorb, go get some water, go use the bathroom, and then we'll come back and I'll show you some really cool stuff that you can kind of do in FLORA besides just kind of connecting nodes and whatnot. Okay.

So let's see here. Let's see. Can you control the degree of difference between multiple images in a generation? Can you control the degree of difference between multiple images within a generation? If you want, so in that case, you would need four different prompts, or four reinterpretations of the prompt.

So, I can quickly just demonstrate that. If you think okay, all these four, like all these four generations that I just hit, for you guys before, actually, if I just undo this here, if I look at these and I'm ah, these are all looking the same. They all have the same black syrup on them. They all have the same green glasses on them. I want four different versions of this, like with different interpretations of colors and whatnot. What I can do instead then is generate four different prompts.

So we're going to demonstrate that here. Select these move over to the left here. So what I can do is actually, once I have this, image node plugged into this text node, that's going to write me a prompt. I can actually do four variations of this text node.

So again, if I hover over this generation button, I can do four. And then now I'm going to get four different prompts, that will, kind of inter like I'm going to get four different texts, generations based off of the initial prompt I wrote. So write a prompt that would make this image attached Halloween themed. One thing though, I have all of these set to the same initial prompt.

So I have this one, purple and green themed, purple and green themed, purple and green themed, purple and green themed. So all of these are going to give me different prompts, but they're all going to be purple and green themed. However, maybe I want. The first one to be purple and green. Maybe I want the next one to be orange and black. Maybe I want the next one to be red and black. And maybe I want the other one to be, black and white.

Now I can just select the ones that I just changed and I can just rerun them. So I'm just going to select those three. I want to keep the first one the same. So I'm just going to select these three. That I just edited and then just hit enter and they'll all regenerate.

So now I'm going to have one prompt that's purple and green themed, one that's orange and black, one that's black and red, and one that's black and white. So I'm going to line these up here. And now what I can do is extend an image node out of one of the text nodes, make sure big, big note here, pay attention. If you're doing this with multiple text nodes, make sure your original image is plugged into that image node, because if not, it's going to give you something different. And I'm going to actually demonstrate that I'm going to extend this text node into an image node. And I'm not going to attach my original pancake image here, just going to hit generate and see what we get.

But in this case, I'm just going to make sure everything's connected. Just extending them like I showed before into an image node from the text node, extending an image node from the text node and plugging in my original image. Just kind of move these around. And then I can select all of them, hit enter on my keyboard.

And then now I'm going to get all those prompts generated at the same time. So that's another way. See, yeah, it's a little different. It's not exactly the same, but that's what happens when you don't plug in the original image into the generation.

So that's kind of how I would go about getting four unique variations out of one prompt. That was a long answer to that question, but I think that's kind of interesting. You can use variations for text nodes as well. Another question, how many reference images are ideal before consistency starts to get worse? Great question. I'm going to go over that as well as elements later on after we come back from the break, which will be in around five or 10 minutes.

But yes, I will bookmark that because it's always a tricky thing. Sometimes people think that the more I feed an AI model, the more references it has. Sometimes that's a good thing. Sometimes it's going to bite you back, and you're going to be struggling with why is it hyper-fixating on that one image out of all the references.

So I'll kind of address that later. Let's see here. Are you able to show us edit tools like in paint? I can quickly show that to you. Sure.

So this looks awesome, damn. I can't wait for Halloween. Let's say I like this image, but I want to replace that dead cherry on top of the pancakes with something else. If I click this image here and I go to tools, you will see a lot of options here. Again, I'm treating this very much as like an entry level type of exploration of the canvas and the nodes and whatnot.

So bear with me here. I know there's a lot going on here, but I'm going to show you in painting since it did come up. I'm going to click in painting here. And again, if you need to look for that, if you select the node that you want to edit, I'm going to go up here to tools at the toolbar.

And then here, it's going to be the fifth option down called in paint. I'm going to go ahead and click that. So now you have this little circle with a plus button and it's perfectly sized to kind of just select that little cherry there. If I want, I could just like kind of click and drag and kind of, take care of like a little bit more surface room. If you need to make the brush smaller over here on the right side, you'll see a few. You'll see like a size option there that will change the size of the brush.

And then you can also pick the resolution of your final generation. I'm going to keep that to 2K, but here I'm going to type in, replace the black. I don't even know if like the AI will understand that's a cherry, but I'm going to try replace the black cherry on top of the pancakes with a candy apple. No, not candy apple with a candy corn.

And then I'll just hit enter. And now that's going to branch off into another image note. That's going to look like this and it's going to be generating. You're going to know that it's generating because the noodle and the border of it is kind of highlighting and animating.

So I'll let that kind of go, but that's kind of how you would go about in painting. Let's see here. I'll answer one more question before we break. Let's see.

Okay. That's super easy. That's, that's a very, I'll take that one and give you guys a freebie. Router nodes are just like you ever, like behind your TV, you have all those cables just kind of sprayed all over the place. And I don't know about you guys. I'm really bad at cable management.

So it's kind of the same thing with noodles. You can kind of see that noodles start to get a little crazy here. Like over here, they get a little crazy. Another example here.

This is a good one. This is kind of crazy. Like the noodles just start to get, a little disorganized router nodes are really good zip ties for noodles. So if I press R on my keyboard, this little thing, this little cute guy, this little router node will pop up.

This is let's say if I want to, if I'm all the way here, right? And I want to generate this image into another image node. And I'm Ooh, damn, I need my original Arizona image here. I'm Oh, I got to extend that all the way here.

And then I'm going to lose track of it. Like that noodle is going all the way there. Instead, I can plug that Arizona image into a router node. And I can just move that a little bit closer to that image node that I want it plugged into.

And then I can plug it in there. It's exactly like an extension cord and or zip tie for your noodles. Kind of helps like if you're ever wondering like what the hell is plugged into this image node, I can just ah, this router node has that image that also leads all the way back here. It's just kind of like to keep your to keep your canvas a little tidy. Otherwise, it has no real function other than just helping you clean up.

Okay, that was a freebie. One more question. Let's see here. Let's see.

Okay. How do you isolate parts of the node? Like how you have been doing? How do you isolate parts of nodes? Unless you're kind of maybe I extend this into an image node and I'm calling out change the cactuses and the image attached to wacky inflatable flailing men. That's one way to kind of isolate, parts of a node. There's another tool that I'll show you after the break called split into layers.

That might also be helpful for you guys. So I'll show you that to you as well. But that might also help answer that question. Ah, yes, there was not the one other thing that I wanted to show you guys, which is kind of to help with a canvas organization.

So I know I showed you guys router node, which was a zip tie or an extension cord for your noodles. But you'll see here. Have these like little gray empty boxes. These are called fields.

So if you're familiar with other design tools, such as like Adobe programs or Figma, you're probably used to grouping multiple things on the canvas together. In this case in FLORA, we have that same thing. So let's say I have this big workflow here. I can just click and drag to select all these nodes.

And then I can right click. And then I can go to Oh, my God. Sorry, you can feel them here on at the top toolbar. Just click that. And that will group those nodes together.

So now if I click and drag this field, all those nodes will move together. If you want to create an empty field like what I have going on in my canvas where I split things out. So my ADHD self could kind of focus on one thing at a time. You can click F. On your keyboard.

And then click and drag. And that'll create just an empty kind of box. You can also color code your fields. So if you hover over a field, you can select the color. Like maybe you want it blue, maybe you want it green to kind of keep things a little bit more visually differentiated.

Okay, let's see here. There's a question about separating or splitting an image into grids. So I'm going to demonstrate that since I don't have an image that is a grid itself, I could just demonstrate it with any image. So let's say let's pretend that this is an image of four different generations in one image node. If you select that image node, and you go up to tools. If I go over here, oh, no, sorry, split into grid.

So it's going to be quite towards the bottom or in the middle. You can now kind of go off and be okay, if you have a image of four image or an image node with four images inside of it, then you want to drag your cursor or you want to drag your mouse so that four boxes are highlighted. So what this will do is if I click with these two, a two by two grid selected, that image will be split. Evenly into four different image nodes.

So that's kind of how I go about that. If you have a image with, let's say, a three, like six images in it at once, then you can kind of click and drag out until you get like a three by three square grid, click that. And then you'll get that as well. Funny enough, I love using this tool, even if I don't have a grid, a gridded image. Because I love actually just getting really cool. I don't know, there's something really artful about that. The spiderwebs, the focus, the close in about that stuff.

If you ever need, if your composition is ever looking bland, try splitting it into a few grids and getting close in croppings of little moments in that image. And you'll get some really cool results out of that, such as this little corner with some spiderwebs. But functionally, if you have an image that generated with four images inside, that is what this tool is meant for. So kind of go back there.

Let's see. Okay, so I'm going to pause there with questions and kind of move on from there. Let's see here. Okay. From here on out for a little bit, I would like people to just kind of pay attention for a little bit, and then just kind of get a sense of the tools that I'm about to show you. It's just so that everyone kind of gets a sense of like where to find the tools, how they function.

And then once we're kind of at the time for independent like studio time, where you can kind of try them out yourself, you can kind of go back to your own canvases. But for about the next 15 to 20 minutes, if you could pay attention while I show you guys these tools, so exactly where to find them in your own canvas, that would be great. If you already know about techniques, video editor, etc. That's cool, too.

But in other words, I'm going to kind of show you a few really cool tools that I think will kind of help expand your workflows from these basic stuff that I the basic stuff that I kind of showed you from here. So I know I'm kind of zoomed out right now. But I'm going to kind of go ahead and focus in here. Oh, one thing, Olivia, if you could grant me a little bit more wiggle room on my canvas, that would be great. What I mean.

So let's start with techniques. So you guys kind of saw me doing quite a few things up here, such as combining these two images so that the glasses sit on the pancake, or maybe combining these two images so that the glasses are made out of the pancake, all that sort of stuff. Or maybe, by the way, that was the generation off of the out painting or in painting. So that was a good example there.

Okay, back to techniques. Maybe you saw me do this workflow where I had this image plugged into five different text notes that are writing five different prompts to give me, oh, sorry, four different text notes to write me four different prompts to give me four different images that are Halloween themed based off of like four different color palettes. I can condense this workflow down into a technique so that any image I input into this workflow will get the same treatment on the other end. So what does that look like? I'm going to show you how to create your own technique. And look, if you don't have a workflow that kind of pertains into that kind of lends itself really well into turning into a technique, don't worry about it. We will have a floor allow an other floor labs where I kind of go more in depth with these tools.

But I'm just going to quickly demonstrate this for you to kind of just kind of show you what's possible in flora more than anything to kind of just like what you're appetite a bit. So I move this over. On the left, I have my input image, which is my image of the glasses on top of the pancakes in the middle. I have those four text notes that I showed you earlier. And on the right, I have these four different outputs. I want to now turn this into a technique because every workflow, every technique is a condensed down workflow, but not every workflow can be a technique. And the only reason why this workflow is set up really well to lend itself to being a technique is because I have a very clear input and I have a very clear output.

So I'm going to go ahead and click build. And the first thing that it's going to ask me when I select technique builder is what are my inputs? So since I already defined before I did anything, what do I want people to plug into this technique? It's going to be an image. I'm going to go ahead and select my image on the left. And I can name this whatever I want. I'm going to make this technique so that you can plug in any image you want.

So I'm going to just say it. And I can write it to subscription. This is where you can input any image. You want to make Halloween themed and kind of go from there, press continue. And now it's going to ask me what are my outputs. I'm going to go ahead and select all the four images on the other end. You see that I'm not selecting any of my text notes because when you run this technique, I don't want you to be seeing any of the text notes.

I don't want you to be seeing any of the text notes. Any of the prompts, I'm trying to make this as easy as possible where you plug and play, plug your image in on the other end, you're going to get four different new images. So that's kind of what's going on here. So now we can name these.

So I'm going to say this first one, which is orange and black. I'm going to name that orange and black. I'm going to name the second one purple and black. I'm going to name this third one black and red.

And then this fourth one black and white. I can add descriptions here as well. I'm just going to keep that the same. I'm going to hit continue. And now I can name this Halloween themed technique. Turn any image you want into a Halloween themed image. I could do a few things here.

I can publish it to the workspace, which I will do. Like that. And then I can also select the category if I wanted to. So in this case, I'll kind of do it just for fun and inspiration. And I could also add some tags, but for the sake of time, I'm just going to go ahead and publish this. And now what you'll get on the canvas is a green note here.

So this is the workflow that we started off with. And this is that same workflow condensed down into this nice little neat green note here. So now I'm going to test run this. I'm going to take my bottle image here that I generated in the beginning of class. I'm going to go ahead and copy that here. Paste.

So I'm just going to plug that in to my technique. And now I'm going to hit run. I don't got to write a single prompt. I don't got to select a single model. I don't got to do anything. Catherine already built this technique for me. She already went through all the trouble five minutes to 15 minutes ago.

It's fine. I just got to plug it in and play and see what I get. So that's a really good kind of example of building your own workflow. Trust me.

This is a little bit of a rabbit hole. So I'll kind of pause. I'll kind of like stop there in regards to building your own custom techniques. But if you want to practice building your own techniques, we will follow up in an email with a few more resources that kind of go over this stuff in more in depth. For today, I'm just going to kind of continue on to show you guys a few other things on the canvas, such as a few techniques that we already have that are ready for you to kind of experiment with.

So on your left hand sidebar here, you're going to see this little icon. It's the third one here. And you're going to click that. And this is going to lead you to your techniques, our technique community page.

This is all just a big collection of techniques that we already built for our community. A few of them are built by users of FLORA. These are workflows that are condensed down into these texts and technique nodes that you already saw me build, but function in a lot of different ways. So if you see one that you really such as color palette extraction, you can kind of click and drag onto the canvas and then it'll pop up on the canvas there ready for you to kind of plug and play.

So a really good example here is the color transfer technique. So if I go back into my techniques community page here and I type in color, it's going to be the color transfer technique here. So you can kind of just click and drag that out and that'll appear there. So here's an example where I took one of my Halloween images and I took my image of the flower, plug those into that technique. And I made sure that I plugged them in accordingly.

So I knew that this Halloween image was going to be my color reference input. And I knew this flower image was going to be my color receiver. So when I had to plug those in, hit run. And now I have that flower image. Halloween image attached. Kind of love that's kind of cute.

So if you guys wanted to try this out again, you can kind of go into your techniques page on the left and then search up the color transfer technique. But don't go just yet. Let me show you a few other things. Another technique I really love. We're back here with our red shirt guy. We're plugging him into a technique called cinematic movie stills.

So this one's a fun one. It's you plugging in an image and then getting four different shots based off of that image on the other side. So now we have my man smiling. Now he's looking a little more mysterious.

Now he's like a little to the side. Now he's like a little bit to the other side. And you didn't have to write a single prompt. You just plug it into this technique and it kind of goes and does the thing for you. They plugged in a very static image. The four outputs are a little bit more kind of like very much in that world.

But test this out as well. Cinematic movie stills technique is really great for just taking one image and turning it into four alternative versions. Very similar, but a little bit more dynamic is the storyboard maker. And I love this one too.

This is plugging in one image into the technique. And getting nine different storyboard images on the other end. So if I actually open this up, so if I hover some of these techniques, you can do this. Some of these you can't.

But for these, if you click a technique and you go to view workflow, you'll see all these text nodes in here that you don't ever have to deal with ever, ever. Don't worry about it. But this is kind of what's going on under the hood to kind of give you a sense of how complex these techniques can get. But all you're seeing is the output on the right.

So what this is doing is nine different prompts that are writing nine different storyboard images based off your input image. So I go back. That's kind of what's going on here. I plugged in my two shoppers in the middle of Arizona. I have no idea where they're shopping. I want to go. It's fine.

But on the other end, we have them. I don't know. Maybe they're at sunset. Maybe it's a closeup of the bags. They're walking and they're kind of like stomping their feet. Maybe they're framed from cactuses, all that sort of stuff.

So with just one run of a storyboard technique, I get nine different perspectives or nine different storyboards from that one image. So incredibly helpful technique. I love this stuff. Really helpful in that sense.

So those are three main techniques I really like to use. Of course, I'm not telling you like you're not allowed to use any other techniques. When we have open studio time in about like 15 minutes, I encourage you guys to kind of experiment. And again, where you can find those techniques is on the left hand side bar over here under techniques. You can click that icon and this will be where you can find all those techniques that are available in our community page. And if you want to search for anything, you have a search bar up here.

OK, so I'm actually going to revisit this technique I made and kind of see how it went. I'm going to move this up a bit. Can I? OK, all right. Oh, sorry, let me just undo that. If you're running a technique and you'll see like I these nodes off of a technique kind of stack vertically, they get a little crazy. Like if this is like I don't want them stacking like this, click and drag those nodes and this works for any nodes on your canvas.

Any nodes on your canvas that you want to reorganize or reformat, just click and drag to select them. And up at the top, you're going to see this option. Oh, sorry, this option here called auto format. If I click that, it's going to just format them so that they're a little bit more tidy into maybe like a two by two or like a four by four, et cetera.

So if you ever think that your nodes are going to kind of get too crazy, you can kind of do that warning. And I'm going to do this right now so you guys can learn from me and I'm going to sacrifice. I'm going to sacrifice a workflow. OK, I'm going to do this one. When I actually copy this and then sacrifice it so it's not really sacrificed. If I select an entire workflow and I auto format. Don't do this, because what's going to happen is things are going to get a little wacky.

Things are going to get a little harder to follow that beautiful flow of nodes and the beautiful flow of like noodles kind of leading into each other, kind of like a breadcrumb trail will get a little bit too condensed. So you're going to start to lose track of things. Don't auto format an entire workflow, just auto format maybe a few nodes off of a technique or maybe a few nodes off of a specific workflow, but not an entire workflow itself. So that's my quick tip. Yeah, but let's take a look at these haunted water bottles. Love it. Honestly, I'm pretty happy with that.

So if you want to try out that technique, I think I have it set published into the workflow or into the workspace. So if I go back to my techniques and I type in I think workspace here. Yeah, you should see a technique called Halloween theme technique and I see someone else published a product shot technique there too. So that's cool.

So yeah, give it a try if you want. You can just kind of click and drag it into your canvas and then plug and play. Okay, so that was techniques. Let me show you video editor.

So I know everything I've been showing you guys as of right now have been like text and text node, image node, etc. Now let me show you a few things with video, because everything I showed you in regards to creating a really good static image also serves really big purposes when it comes to setting up for video generations. So why? If you want a very good video generation, you are probably going to get like 90% of the time if you have a still image to work off of your generations will be a lot more accurate. They'll be easier to kind of follow. You'll be a lot more. They're kind of more predictable in that sense.

And you're just kind of feeding the video node exactly what you need, what it needs to give you what you want. So really good example here is if you're trying to generate a video. Personally, every time I generate a video, I'm using a text node. Me, by practice, I'm a graphic designer. I have never done any video work or photo work before dabbling into AI.

So I need help writing prompts to animate things. So that's why I always use text nodes to help me do that. If how to move the camera from left to right, in a cinematic still like I can't even name I can't even name cinematic like camera movements off the top of my head that tells you a lot. That's why I use text nodes to kind of help me do that.

Okay, anyway, let's say I have this image here. And I want to animate it. I don't know, I think I have a sense of what I want. But I'm going to use the text node to help me write that prompt.

So I'm going to go ahead and plug that into a text node. And you guys saw me do that earlier, but I'll go ahead and demonstrate that one more time here. I have this image. And if I want to click and drag out, I'm going to let go and select text. And now I'm going to type something in write me a prompt that will add subtle motion to this image. I want it slow motion. To find nothing too complicated.

I don't know. Yeah. Maybe they're walking or question marks. It just goes to show like truly just like board vomit into a text node. It'll kind of get a sense of what you want.

So that text node kind of hits the round running gives me a really good prompt. And now I'm ready to plug this image node into a video node. So I'm going to click that. And I'm going to go ahead and leave that in the same aspect ratio as that image. And now you guys saw me do this 10 million times before earlier.

So I'm going to make sure that my text node is plugged into that video node. And you'll see that the prompt that was just written for me is going to be populated in that video node here. Now, I think for video generations, just so you don't burn through your usage in the workspace, I would recommend using Minimax H3 Max and make sure that you're about either five seconds duration and like maybe 768 P. That's a very specific number, but that's fine.

And then you would just kind of once you have those set up, you can just kind of hit generate there. So that's kind of how I go about generating videos based off of a text node. That's writing me the prompt. And an image node as well.

So I did that a few times here, as you can see, I did that four times. So I have four videos. So again, just kind of playing these videos here. Great. My people are walking. They're we just spent, I don't know how much I'm shopping in Arizona. Great.

And now we got this image kind of animated as well. Now they're kind of relaxing. And now we got a bird's eye view of these guys walking. And now I'm okay, I think I want to kind of stitch them together. Well, what's nice is that we have a timeline editor in FLORA now.

So if I actually select all four of these videos and I click and drag out and I let go, you're going to see an option here called timeline. You can go ahead and click that. And now if I hit edit in the toolbar, look at that. I have my own little video editing interface here.

So I'm going to just meet these on another thing. You also have audio control for your videos. I'm not going to play audio in this demonstration. I'm going to save everyone the trouble of troubleshooting audio. I'm just not going to do it, but you do have the audio does carry from your videos into your timeline editor.

So if you had this video selected, you can also go down to the right hand side where you see audio here and you can control things like audio or volume. Sorry. You can control the fade in, the fade out, et cetera. And you can do that for any one of these videos, but while I'm in here, I can be okay, maybe, I don't want to start off with the guys kind of like walking like that. I think I want to move up there. I think I want to start off with the guys kind of like walking like that. Move that up here, move up there.

And no, I want that last. So you can kind of start messing around with how these clips sort of interact. And another really fun thing, there's kind of small here. So I can't quite zoom in, but when you hover over a video within timeline editor, you're going to see this white little box there at the top, top, right corner. If you click and drag that, you're going to activate the little fade out option.

So now you can kind of, okay, I wanted to fade at 0.4 seconds. So now if I go ahead and play this, Oh, I'm going to hit that. You can see that it's going to kind of fade in instead of hard cutting to that video. So yeah, a video editor is great for just kind of stitching things together to kind of just see things come together a little bit easier. And so you don't have to kind of generate a 30 second video and hope that it comes out good. You can just kind of generate multiple smaller short clips and then put them together in video editor.

So a quick question in the chat, where did the audio come from? In this case, I use Seedance 2.0 reference. Seedance just automatically generates audio for you when you generate videos in the model. So there's that. I won't be going over audio node in this demonstration, but we do have some other reasons. We have some resources on our YouTube where we go over audio note as well. Great.

So that was video editor. So if you want to add motion to your haunted pancakes or your haunted pumpkins, feel free to do that. Okay. So that was video editor. We went over techniques. I am going to move on to layer editor.

So this is one of my favorite tools. Shout out to any who's one of our engineers that kind of spearheaded trying to get this together and improving it every single day. Layer editor is like a mini compositing tool in the canvas. So I have two things going on here. I have my pancake image, my beautiful pancakes, and I have my flower image. I want to make a composition. Where the flower is growing out of the pancakes.

So to do that, I am going to actually connect this flower image. I'm going to demonstrate this here. Whenever I'm prepping something for a layer editor, I'm kind of deciding a few things. I'm kind of deciding what is my base image and what is the image I'm going to be layering on top.

So since I know going into this, I want the flower to be layered on top of the pancakes. I want the flower to be layered on top of the pancakes. So once I do that, you'll get a transparent image of the flower. So now I'm ready to kind of move on.

So I'm going to go up to tools. I'm going to go up to tools. And then if I go here, there's an option to remove background. I'm going to go ahead and click that. And that is, ooh, I'm out of usage. It's fine. I already did it.

Once I do that, you'll get a transparent image of the flower. So I'm going to take my pancake image and my flower image. And I'm going to go ahead and select those two images. I'm going to click and drag into layer. Oh, sorry. Let me do that one more time. I'm going to select these two images, click and drag out.

And on the bottom here, you're going to see layer. That is what you want to select to get access to the layer editor node. So if I go ahead and select that now, you'll see here. You have your pancake image or whatever image you want.

And then if you move it around, oh, no, there's the flower kind of hidden underneath. So I can do a few things here. I can select the layer editor node and say, I think I want it two by three aspect ratio. Actually, you can do that. I'm going to actually undo that since it kind of squished it a bit. Shouldn't usually do that, but you can change the aspect ratio of the layer editor as well.

But here, if I want the pancakes to be full screen, I can just kind of double click that and it'll kind of show up there. And over here on the right hand side, you'll see that I can change the arrangement of the layers very similar to other kind of like layer compositing tools. And I can put the pancakes all the way at the bottom so that it's in the back and that the flower, my little itty bitty little red flower could just be placed there. So I can just kind of place this. I have layer editor allows me to have a lot more pinpoint control over okay, no, I want this flower to be huge and I want it to be growing out of the pancakes like that. Or I want my pink. I want my flower to be super tiny.

And maybe I want it on the side rather than the center. So layer editor kind of helps you like take control of that. I got lucky here. And like these, this flower looks like almost beautifully like incorporated into the image.

But I'm actually going to remove this. And here's an image where I have the pancake and the flower kind of attached here, I can control it where it is. And so now once I have that layer editor kind of set up the way that I like it, I can go ahead and click and drag that layer editor out into an image node. And I can prompt something like incorporate the flower in this image seamlessly into the pancakes to make it look like it's growing out of the pancakes. And so that's kind of what's going on here.

This layer editor that we just kind of went off of is serving as a rough sketch or rough composite that the image model on the right will now kind of be okay, I'll, I can make that a little bit more realistic. I can make it look like the flower is actually growing out instead of looking like it just pasted on. So that's kind of what's going on here. So layer editor is really great for that giving you finer to control over how big a thing is in relation to another object, having a little bit more custom control over how things are looking, so that you leave a little bit less guesswork for the AI model to kind of interpret. Another really good example of layer editor is you can also add text into it.

So here, I took one of my kind of haunted pancake glasses at images. And I'm going to plug that into another layer editor again. So again, if you click and drag out of an image node, or multiple image nodes, you let go. And on the bottom here, you'll see layer. Alternatively, if you go to your left hand sidebar and go to add node, you'll see the layer option there as well.

So now if I go back into my layer editor, I'm going to go ahead and double click that so it goes full screen. And I can add text in this now. So if I activate my layer editor, and I go into my right hand toolbar, I have an option to add text. So I'm going to click Add Text. And now I can kind of okay, maybe I want my headline to go here. And what did I write here? I said, ghoulish glasses.

I don't know if I wrote that right, but I'm human. It's fine. Ghoulish glasses for every occasion. I'm pretty sure occasion has like two S's in it's fine.

So now I can kind of like control where that text is, how big it is, all that stuff. So let's say I like that. Maybe I want it a little bigger, maybe 180 in text size. I can control the things like letting.

So letting is the space between each line of text. So if I just kind of downsize that there, you'll see that the lighting is a little bit tighter. And then I can also control tracking or tracking which is how close the letters are to each other. So if I want them to be a little bit tighter, I can do that as well.

So we have a few little for my typography people out there, we got a little a few things in here for you. But the biggest one is if you have a custom font that you want to use, that you just want to quickly visualize on this generated image, you can do that too. So if you go into typography, and you go into your font selection here, on the bottom, you'll see upload custom font. And if you click that, it'll open up your finder on your computer, and you can then select whatever font you want.

So here I have a few I have yowza, which is, I forgot the type foundry's name, I have buseet, you can upload a lot of different kind of custom fonts that you may have on your computer. So here's an example. I have this layer editor where I have two text bodies here, I have one for my headline, set in a custom font, I have one for smaller text here. And here, you'll see a little circle shape. That is here next to the text option on the bottom right, there is a add shape here.

So if I click that, you have an option for a square, a circle, or triangle. I think I just popped up, nope, definitely not a square. I popped up a circle, and I just kind of transformed it a bit to make a little elliptical. A condensed circle, and I just colored it pink.

So if you want to change the color of that shape, you can kind of go up here to the swatch here, and you have a little color selector there. If you have a specific hex code, you can also paste that in here too. So yeah, another really cool thing you'll see here, I have a black square. If I hide this, you'll see that the text gets a little hard to read. I have that black box there, so that the text is a little easier to read. That's a trick that I use all my career. I'm sure you guys are familiar with doing that as well.

But if you wanted to do that, you'll see there's an option here for opacity. So if I set that all the way up to 100, then it's totally black. If I set it to about 50%, there you go, then the text is a little easier to read. So all of a sudden, I went from like my, okay, kind of fine haunted pancakes to come into my store and buy my glasses for Halloween. And that looks like 150% off to use the code spook26. That's hilarious.

But you can make a really quick mockup of what a digital ad might look like with a generated image straight into FLORA, just to kind of quickly test the vibes, kind of like how it's kind of feeling in the context to custom typography, colors, all that sort of stuff. I'm going to show you my other favorite technique. I didn't show this to you before because I needed a really good example, and this is a really good one. Let's say you just spent the last five minutes just kind of pulling this together. I'm going to plug that layer editor into a technique called layout resizer.

So if I go back to my technique window here on the left side, and I type in aspect ratio, the second option called layout resizer, is going to let you plug in any layout you want and get five different aspect ratios on the other end. So if I click and drag that onto the canvas and I plug my layer editor into this layout resizer technique, on the other end, I have that same layout in a 16 by nine, in a, what is this? Whatever aspect ratio that is, I think it's, there you go, four by three, nine by 16. 21 by six, I think, and then this one as well. So I spent five minutes on a layout, and then now I have five different aspect ratios based off of that same design. Really, really helpful to just kind of quickly visualize what a layout might look like in different aspect ratios.

So maybe I just got to see what this looks like in an iPhone. We got to present this to marketing in like five minutes and my boss is asking me, Catherine, I really need to see what this ad will look like in an iPhone. You can kind of click and drag that image, that output from the technique into another image node. And the same thing that we've been doing since the beginning. Just kind of prompt it as simply as I can say it. Mock the attached ad as a social media story ad on a black iPhone 16 Pro. It could be an Android, whatever it may be.

No hand, white plain background. And now I have that mocked up for me on an iPhone. You can also do this for something like maybe you have an interface like a website page that you need to mock up the spanner on. You just need to quickly kind of see what that might look like.

So I have one of these wider aspect ratios plugged into an image node saying mock up the attached ad as a digital ad on a sunglasses website interface. Of course, if you have an existing interface that you want to mock this up on, definitely do that as well. You don't have to generate your own. But yeah, that's a great way to kind of just see things in context to where they might actually live.

So yeah, layout resizer is a really great technique. One other thing, we have the same, a very similar, we have two versions of this technique. If you type an aspect ratio into the technique page, you'll see two things, photography, image, aspect ratio change, layout resizer. These two techniques are built differently so that if you are plugging in a photograph, then the prompts are specifically written to resize that photograph as images. If you are plugging in a layout, such as the one that I plugged in here with typography, shapes, et cetera, then layout resizer is better for that use case.

So when in doubt, kind of pick and choose which one will best serve you for your particular workflow. Another really great example of layer editor is, again, kind of similar to what I did with those haunted pancakes, except I'm taking my shoppers again, and I'm making a nice simple brochure here. You'll see here I have a transparent image of a logo. It's just something to kind of plug into the layer editor with my photo. And again, if I click and drag and select both of these, I can just select layer on the bottom, and they both pop up here.

So I can just double click this, place that logo there, et cetera. But what I did here was I just added that logo at the top right left corner, I added some text, and then I plugged that layer editor into an image node. And my mockup people out there that are battling with smart objects and linking files all day long, I feel you, pay attention to this. You can make your own mockups now, you can breathe a sigh of relief. You can just kind of plug that layer editor into a image node and just say mock the attached composition up on a brochure cover against a white plain surface.

And then you got something like that. I took it to another level, I kind of plugged that brochure into another image node, and I said, ooh, change that top left icon to an embossed gold foil. Damn, that's fancy. I want to go there. Community with a nice golden icon. Yeah, take me.

So that's really fun to kind of experiment with. I did another version of this where I zoomed into that logo. So in this layer editor, I attached this brochure cover into another layer editor, and I just have this that image, like really sized up. I just dragged out like around there to kind of create a little bit more of a dynamic kind of focused in composition. Plug that layer editor into another image node.

And then I just said, render the symbol in the image attached as a raised gold embossing. And I get something like that. So if I had this like two years ago, I'd be losing my mind right now. So this is honestly I'm just so happy I can kind of do this without like wrangling perfectly lighting and embossing like a gold texture in Photoshop. It's just so much easier now.

So yeah, I love layer editor for that stuff to kind of just get a rough sense of a composition so that an AI model can kind of take it from there. Okay, two more things. And then I'm going to let you guys kind of just like play around and then you can either watch me live build stuff, or I'll kind of be snooping in people's canvases until the end. But I have two more things to show you 3D node and actions.

So we're almost done here. But 3D node is really fun. It's a new release that we just kind of came out with. But you can take any one of your images that portray a singular subject, such as for example, here, I have my stack of pancakes.

Now I'm oh, I want to turn this into a 3D object, because I want to change around the angle of this image. And alternatively, I could just kind of go down a prompting hellhole and be give me this image, but with a three quarter view. Now move it down. No, more down. No, too far down. Move it up, blah, blah. It's fine.

What? No, 3D node will take care of that for me. So I have my image of my pancakes. I'm going to demonstrate this again, brush on the left side. I can kind of click and drag out. And I'm going to go ahead and select 3D node here.

So a few things. How much is going to kind of pop up at first, you don't have to prompt anything. There are a few options here, depending on the model you pick. Don't worry about picking the model. Just go for meshy, anything in meshy. Go for meshy seven. If you're going to do any 3D stuff, and you're just not you've never kind of worked with 3D models before.

I know I haven't. So I just go, okay, meshy seven sounds good. Let's just go with that. You can kind of just pick that model. I'm not I'm not going to mess with any parameters because I don't know what any of this does. I'm just going to go. I'm just going to click run.

And what I get on the other end is a 3D node. I could spin around these stack. It even rendered the bottom of the pancakes. It's pretty crazy, honestly.

So I can kind of click and drag to kind of take control of the camera angle. To really observe how do these how do these pancakes look right now? Let me let's let's like observe these crusty edges, huh? Okay, not bad.

So what's nice is that I can kind of okay, I kind of move my pancakes there. And I kind of want to just like drag them into corner there. I can kind of set up any camera angle I want. And you'll see a few options here at the top left corner. You see photo, video, and then sheet. I'm going to take a photo.

So what this is doing is just kind of taking a screenshot of that 3D model at that exact angle. So if I click this, give it a second. It's going to go ahead and generate a transparent image of those pan of whatever you have. It doesn't always have to be a pancake in that composition in that same angle.

So this kind of example of what I did here. And so you're probably thinking okay, well, Catherine, that looks kind of like a what do I do now? Well, if you look at this image and you're I want to change the anger, the angle of the pancakes, you do that by creating, turning it into 3D node or 3D object. You take a picture of exactly the angle you want.

Now you're going to take a picture of the pancakes from there, and you're going to plug that into an image node with your original with your original images or your original image. And then you're going to prompt something like re-render the first image, which is your picture of the 3D node. So re-render the first image in the same style as the second image, which is our original pancake image. Keep the first image angle and background color the same.

So that's kind of what we got. So if we compare, I'm going to put these side by side here. I don't even know how I would even begin to prompt an image node or a text node. I don't even know how to begin to name this camera angle. I just don't know. I don't know what you would call that, like three quarter view of the pancakes in the lower right corner, something like that.

But what's nice is that doing this kind of thing gives you way more pinpointed control over how a composition can kind of look, and then you can re-render it in whatever way you want. So another example here, I plugged my glasses image into a 3D node, angled it in a particular way, took a picture, and then I did the same thing. Re-rendered it with that original image plugged in. It's really helpful for that as well. Another example here is I took that image of that logo. Oh, I just missed that.

OK, here. If you have to make the 3D asset, all you have to do is plug your image. And here, I'll do this over here. Plug whatever image you want. Ideally, it should have a focal piece, like an object or something.

And then if you click and drag out and you select 3D, that's how you get this 3D node. It looks like a normal text prompt kind of on its own. Again, just pick meshy. 7, it's fine. And then you can kind of click Generate, and it'll go ahead and generate the 3D object for you, which is kind of what you see here and kind of what you see here.

So again, did the same thing for this logo. I took my logo, plugged it into a 3D node, and now I have a 3D rendering of that logo. So now if I need to turn this into environmental signage, into wayfinding, into a sculpture for a community, all that sort of stuff, you can do that as well. This is a fun one. Actually, I didn't check on this generation, so let's see how it came out.

But let's say I have this gummy bear here, and I'm going to show you guys how to record a particular 3D movement. So I saw a question in the chat. Can you use 3D to make a 360 rotation video? Yes, you can. If I click video, you're going to see a little record button pop up.

So now I'm going to set this up first. Hold on. I'm going to start off here. I'm going to click the red button, and then I'm going to go ahead and just slowly rotate until I get around 360 like that. Hit record or hit stop.

And then what you get once you hit stop is a recording of that object. Moving in 360. So I did one like that. I did one before. And again, since I kind of suck at writing prompts, and also if you notice, this rotation is a little jittery because I'm literally rotating it with my mouse. I'm trying to generate that into a video node. That's going to smooth that out for me.

This is actually something I was trying to figure out, so I wonder if it worked. Let me see here. It kind of worked. It's a little jittery.

So TBD on that. It's a little work in progress, but what I did was I plugged that video that came out of that 3D node into a text node. And I said the video attaches a rough 360 rotation of a bare object. The motion is jittery and uneven. I need to re-render this as a smooth professional 360 video of the object. Against a white studio background in a photo realistic way. Write me a prompt that will do that.

So that's an example of me, again, using a text node to kind of help me write a prompt that will get there. One second. OK, so honestly, that's not too bad. I think in this case, since I used Minimax, it did the best it could, but I'll definitely retry that with a better model later.

OK, so another thing that's really cool is let's say I have a sketch of a coffee pot, and I have two materials here. So as you can see, these are flat images. These are not rotational 3D objects. This is just an image of a wood texture and an image of a mosaic texture. I have all three of these. Plugged into an image node.

And then, oh, yes, sorry. So you do have an option to export the 3D asset. So if I go back to my pancakes, if I click, and this goes for any node on the canvas as well, you'll see a download option up here. And for 3D, you have options to kind of download it as a GLB, as an object, and a few other things.

So you can definitely save a generated 3D object from FLORA. So I'm just kind of just plugging these into a image node. And I'm just saying, render the sketch in the first image using the materials in the second and third image. And then I get something like this. And it filters that sketch into a actual object. And I'm OK, I randomly picked those materials out. And honestly, that looks-- that's a good-- that's a handsome pot.

That's a handsome coffee pot right there. Now, let's turn this into a 3D object. So I plugged that generation into a 3D node. And again, just in case, I'm going to go ahead and plug this into-- I'm going to click and drag and go to 3D. I have mine automatically selected to meshy7.

So I'm going to go ahead and just click Generate. And then what I get on the other end is this. So I kind of just-- this is a really good example of apply your materials as flat images. Get a really good-- just like with video generations, getting an image as close to something as you want and then turning it into a video you can do the same kind of principle with-- what do you call it? With 3D nodes.

So here, I just kind of spent the time to render or to generate an image of the coffee pot with these two materials. And now that I'm happy with that, I can now generate that into a 3D object here. So once I have that, now I can demonstrate the sheet option here. So I showed you the photo. I showed you the video. And I'm going to show you the sheet option.

So from here, there are a few things here. I personally-- I'm not familiar with all of them. So I'm just going to focus on the 4-way, the 8-way, and the 16-way. So these are just going to take either four pictures of the object from four different angles or eight different angles or 16 different angles.

So I can go ahead and click 4-way. And if I click that-- it's going to capture four different views of that coffee pot. Now, earlier, I heard someone say I accidentally made an element. So I'll quickly, just super quickly, kind of show you what elements are.

So let's say you're I want to-- this coffee pot is my star. This is the star of my project. I need this in every single place. So I can go ahead and click all of these that were just generated from that four up or that four view option. And again, when you go into your 3D node here of that object, this third icon at the top left corner, sheet, just clicking 4 or 8-way. And either four different angles of that object or eight different angles of that object or 16. Will kind of pop up like these.

So once I have those four, eight, or 16, I can click and drag. Actually, you can't do 16 images in a element. I think it goes up to eight. So either do four or eight if you intend to turn this into an element. I can just kind of click and drag and select all these. Right click. And you're going to see element here.

If I click that, you're going to see this window. And you can name your element. I don't know, whatever you want. In this case, I'd name it Mosaic Coffee Pot.

But overall, what are elements? Elements generate a curated reference for your images. Or they just kind of drive consistency through your generations. So instead of dragging, you can group these four images into a field and then drag them over to the other side of your canvas and then drag them into another video node and be like the four attached images. Use the four attached images as a reference.

But it's kind of like a little messy. Elements help just kind of condense those images into one reference node. So think of it as almost like a reference node. We call them elements.

So you can name these elements like Coffee Pot. You can select-- I wrote that wrong. Coffee. You can select a category.

So in this case, this is more like a product here. And then you can write any additional details about this. So if you want to create an image as this Coffee Pot has a mosaic texture with a wooden handle, blah, blah, all that sort of stuff. It's optional. You don't have to.

But it's there if you want. And then you can either generate-- it'll automatically generate a thumbnail for you. I tend to just like I'm OK with a four-up like this. So I'm going to go ahead and click generate-- or sorry, create. And now I have this one node that's carrying all four of these images here. I can drag this all the way over here.

And then I can click this element out into an image node. And I can prompt something like Coffee Pot hashtag. Oh, sorry, another thing. You see here that I have this element plugged into my image node. If I did not, I can go straight into an image node and do-- if you want to call an element in a image node video node text node, just press hashtag. And that will automatically call up your entire element library.

So here you'll automatically see there's a product folder that I made. And I can just go ahead and click Coffee Pot. And you'll see there that node connection was made for me. So now I can say Coffee Pot on a modern kitchen counter. And just hit go. I'll add some usage after I kind of finished demonstrating this.

So that's kind of how you could use elements to just kind of portray an object a little bit more accurately throughout image, videos, or text as well. Sorry, not text. Elements do not plug into text nodes. Elements plug into image and video nodes.

So that's how they can kind of be used there. So let's see here. OK, last thing I'll show you guys before we kind of go back into our independent canvases. And just start playing around with these tools until the end of class is actions. They're pretty straightforward. If you go to your left corner and go to Add Node, there is an option down here called Action.

So you can click that. And you have a whole different array of options here. If you're not sure which one to pick, actually, if you hover over either a video node or a text node or an image node-- in this case, I have my image node. You'll see this option called Tools. And you're going to see a few built-out tool options already here for you.

So you have a color grade image, an overlay image, a draw, all these sort of stuff. But yeah, these are actions. So I have, for example, this image plugged into a color grade node. So if I go in here and I double click, you'll see that you have things like you can control the warmth of this image. You can control the contrast. Saturation, all that sort of stuff. And these are non-generative processes.

So these are not being generated. These are not using an AI model to do this. These are deterministic actions. So anything you're used to doing in something like Photoshop is what's going on in your action node. And again, how to get to your actions is -- there are two different ways. If you have an image node and you go to Tools, you'll get a few options for image-based actions.

So like color grade, overlay image. You can also go to All Image Actions and get a few selections from there. Or go to your left sidebar and go down here to Action. And then you can select any from here. You can kind of just click and drag. I'll quickly touch on this. I won't get too crazy about it.

But you can also generate your own actions. So if there's a particular effect or if there's a particular kind of filter you want on an image or video, but for some reason you can't find it, you can create it yourself. So I have my custom action node here. So actually, if I repop that up, I'm going to go back to Add Node, go to Action, and then go to Generate Action. And I'm going to go to Action, up here at the top. If I click that, I'll get this little purple node here. And I'm actually going to type-- because I was actually going to create this earlier, but I'm going to say a Film Noise Filter Effect, something like that.

Because I think I wanted to add a little bit of film grain or film noise to an image, but I couldn't quite find an action that already did that for me. So I'm just going to make my own. How when you go into Photoshop and you go into a filter gallery and you see all these crappy filters, like glass mosaic or color pencil sketch, and they're not that great. Creating your own custom action is almost like you're creating your own Photoshop filter. That makes sense.

But except it's really good. It's like creative coding, your own little creative tool. So here, I just generated a film noise action. So now, if I actually-- I'm going to grab-- let's see. I'm going to grab one of my haunted water bottles. I like this red and black one.

So I'm going to copy that, drag it down here. And I'm just going to plug this image into that action that I just generated. Oh, that looks awesome. Look at that's looking spooky. The film grain just adds another layer to it.

So I think automatically, when I generated that custom action, I didn't ask it for grain amount. I didn't ask it for grain size. I didn't ask it for vignette. It automatically just generated those parameters for me to control.

So now, if I experiment with Scratch-- oh, that's interesting. It's just adding random lines. We're going to put that down there. Let me just zoom out here. There you go. Grain amount, that's kind of crazy. I think a little more refined.

So I want the grain size to be larger. Now, I want it a little bit smaller. Vignette? Yeah, I want a little bit of vignette. I'm going to fade a little bit darker. Yeah. Color shift?

It's kind of fun. You generate your own custom action, and all of a sudden, you're in your own little creative program. Again, this is one of those things where you can really go down a rabbit hole. So I'll stop here with custom actions.

But we will send a few other resources about this tool as well, so you can kind of experiment on your own too. Noisied? Yeah, that's kind of cool. So after I'm done kind of messing around, I can just hit Escape on my keyboard. Just Escape.

And then this little arrow here will just kind of pop out that for me. So again, since this is not-- the AI model is writing the code for me that is creating this custom tool. But this image on the right that I just exported or that I just kind of outputted is not technically AI generated. That grain was programmatically or deterministically kind of done on that image.

So no image degradation is happening there. I see a question here. Oh, DJ's going to answer. But you can't save actions as of right now.

But that is something we're working on so that just like techniques, you can kind of save your actions and kind of have a little library of them. So I showed you quite a few things over the last 30 minutes. I got a little too excited about some of these features. But honestly, they're really cool. I showed you again in the beginning. What is a blue square and a green circle kind of equate to? And you saw earlier with my presentation, it can literally give you anything.

But I think just kind of getting you guys set up on the sense of how to connect things and just kind of get a general sense of how things can kind of influence each other on the canvas really sets you up to kind of understand really good use cases for techniques setting you up for creating your own techniques sets you up for creating videos and stitching those videos together with timeline editor, even using layer editor, and then kind of building off of these workflows. So again, the real beauty of kind of like this node based canvases. While one thing is generating, you can go somewhere else on your canvas and generate something else. You can riff off of five different directions at once, all that sort of stuff.

So over the next like 25 minutes, or like I'll say like over the next 15 minutes, I'm going to be here just kind of like looking at people's canvases and also just kind of answer keeping an eye on the chat and some questions. So if you guys are kind of finding yourselves a little bit lost at any point, feel free to drop them in the Q&A or the ideally the Q&A so we can kind of keep track of the questions. And I'll answer as many as I can. So let's see here. I see that there's a question in the chat in the Q&A. I think I went over missed the part where they made 3D asset again, just quickly. If you have an image here, again, click out, and then you can select 3D.

And then don't worry about prompting anything. If this is new to you, just hit enter or hit the arrow here. And it'll generate that 3D asset for you there. So yeah, I would love to kind of for the next 15 minutes, just kind of play around with the tools I just showed you guys. Be sure to stay at the end because I will kind of just like round this all out and close out with a few regards.

But yeah, just going to kind of keep an eye on the chat and the Q&A. I'm going to refresh my canvas too so I can generate some stuff. Let's see. OK, so if you have an existing website, could we use a URL to use as our mock-up or would we need a screenshot?

So definitely take a screenshot of your website. Since you can't technically enter a URL into the canvas, I do that all the time. Just kind of take a screenshot of a website and then just bring it in as an image node or as an image in general. You can open up your file manager or like your computer files, click and drag a bunch of files onto the canvas, and that's one way to do it.

So yeah, you can just kind of click and drag anything you want onto the canvas, including images, videos up to 25 megabytes. You can also upload PDFs onto the canvas as well. So if you have a project brief, that exists in a PDF format and you just need to generate images based off of that information in the PDF, you can do that as well. Let's see here. What is the limitation and resolution for output images? It depends. As you can see, like Nano Banana Pro has, let's see here.

Nano Banana Pro has up to 4K. You create an image in 4K and then you use an uprising model to double the resolution. You will technically have an 8K resolution image. I think the last time I tried that, I got an image that had like about 16,000 pixels in height.

So that's definitely an option if you're printing a billboard or a huge poster. So yeah, I think that's, it can kind of go double or triple 4K resolution. Let's see. I actually had a few things that I wanted to kind of try out on my own on the canvas while I let everyone else kind of play around for the next 10 minutes or so. I wanted to create, actually, can I generate this now? Let me see something here. I'm also going to be peeking into people's canvases soon.

So that's going to be really fun. I'm going to generate that coffee pot on a modern kitchen counter. And another thing I was trying to test out was actually a brochure booklet 3D model. So I can make like more dynamic kind of brochure mockups.

So let me try that out, actually. I'm going to plug this into 3D node. I'm actually going to try and prompt. I've never done this before, so let's see. 3D model of a small brochure booklet.

Let's see what that gives us. Let's see here. So one thing we did kind of skip over was our in-canvas agent called Fauna. So if you look on your canvas, on your lower right corner, you're going to see a few things here. You're going to see a Q, a Generation Q, and also this little circle icon here.

So if I click that, it'll pop up a chat window with our in-canvas agent called Fauna can help you create workflows. So if you select an image for example, if I selected, let's see here. If I selected this sketch of the coffee pot, I can ask Fauna render the sketch or create a workflow that would render the selected sketch as a real coffee pot, something like that. So if I actually kind of go ahead and ask Fauna to do this for me, it'll create the workflow for me.

So like my hands are not on the canvas right now. It'll kind of go ahead and pop up the nodes on its own. Pretty much like it's a really good balance between like a chat interface and the kind of canvas node-based interface. So it's definitely, you can give it a shot and see if it helps you out when you're stuck in a moment. I tend to just kind of troubleshoot things by hand.

So that's kind of my preference. But definitely, if you're curious to kind of just like ask Fauna to do something for you, instead of like picking the model that you want, or picking the prompt that you want, try to see if Fauna can help and kind of go from there. So yeah, yes, Fauna does use tokens. So that's something to keep in mind.

Let's see here. Can you go back to how to select a specific element or a specific part in an image to develop? Yes, I definitely can. So yeah, so Fauna just kind of rendered that coffee pot for me. I think it picked Natto Banana 2 as the model, wrote the prompt for itself and just kind of popped that up. On the node, on the canvas.

So really helpful sometimes. So yeah, definitely if you're curious, give it a shot. But to answer the other question, I'm going to go back to this one here. Okay.

So I'm gonna just kind of copy and paste this one here. So again, to like edit an existing image on the canvas by prompting, you can just kind of click and drag this image out into another image node. And you'll have to kind of, if I'm looking at this image and I'm I want that upper left icon to be gold embossed, I can just kind of prompt that as easily as I can say it, I can prompt it. So I would write exactly that, which is kind of what's going on here.

So I said, change the top left icon to an embossed gold foil. And here I'm not saying like make it gold embossed because no one will know like what I mean by gold embossed. I'm specifically saying change the top left corner to gold embossed. Oh, someone in the chat asked, how do you remove noodles? If you just select, click and drag and select the noodles and you hit delete, that's how you delete noodles.

So yeah, you can delete noodles as well if you have to, just by clicking and dragging and selecting them. Let's see here. So what was the name of the technique that I showed where you can create different views from all sides of an object? So actually, there is a technique for that called character lock. And actually, if I do have an image of a, here you go. I'm back for you. Red shirt guy.

If you have an image of a person, we do have a technique called character lock. So again, if I go to my left hand side bar, there is a technique here called character lock. Just going to kind of click and drag that. You can plug in an image into this character lock technique. And if you hit generate, you're going to get that same person in all these different angles. I'm going to let that generate, but if I view workflow, you'll get a view of that person by the side, three quarter back view and full body as well.

But the other thing that I showed was down here. When you create a 3D object out of an image in this 3D node, you'll see a sheet option here. And you'll see a four way eight way or 16 way. If you click any one of these, you will get four different angles. Of that 3D object. Can a character become an element? Yes, it can.

So once you get those character lock outputs from that technique, you can again select all of them, right click, and then generate the element.