Zapier and ChatGPT For Bannerbear AI Images | Tutorial

Let's build with Zapier and AI (100+ videos)📅 2023-04-07

Watch on YouTube ↗

Transcript

[00:00] Hey, and welcome to today's video on ChatGPT and how to integrate it in your daily life. We're going to be looking at today essentially going over Banner Bear and integrating that with ChatGPT is going to get really interesting here. We're going to go over AI images, GPT prompts and everything that's involved with understanding how to get started with Banner Bear and essentially integrate Chat ChatGPT. You could use this in your social media schedule for personal use and so on. All right, to get started, go ahead and click the link in the description. You're going to be able to create a free Banner Bear account with no credit card required.

[00:30] And we can go ahead and see how to integrate this with Zapier and ChatGPT and get a full idea of how we can take advantage of Banner Bear's API when it comes to image creation and video creation. When you go ahead and click that link, go ahead and click Get started for free. No credit card use is required. You're going to be given access to some free API tokens to allow you to play around and get really an understanding of how to use the software. And as you see here, a lot of the image data that's going to be pushed forward is going to be using AWS, as you'll see when we grab the links and use it for different use

[01:02] cases. So in this context, in today's video, we're going to be using it to make a tweet. From here, you're going to be able to see that there's going to be a lot of interactiveness with developers, and you're going to be able to essentially integrate Banner Bear with different APIs with Zapier and different forms of artificial intelligence. Without further ado, go ahead and create that account and let's jump into today's tutorial. All right, to start off, let's go ahead and get an understanding of ChatGPT. As you know, we can use ChatGPT for social media. So we're just going to go ahead and do one quick prompt here and then we'll go ahead and jump into the process of creating our Banner Bear account and then integrating that

[01:33] of Zapier and then going over a full AI automation that you can start using today in your different flows. So to start off, let's go ahead and give ChatGPT a little bit of context. We're going to be using model ChatGPT for today, which is a really complex model, and you're going to get way more comprehensive answers than you would have not been able to get with 3.5. So we're going to do context. You are a social media manager, and then we're going to say generate one tweet around the topic of a business called

[02:10] Joe's Coffee and a new product of coffee beans. So, as we know, ChatGPT, you can talk to it as it's a human language processor. This is just giving you an example of a specific prompt. But what we want to do here to minimize the amount of labor that we have to do associated to these kind of tasks is we're going to go ahead and use this type of API, and we're going to make it all automatic so we can automatically create a social media photo that can therefore be posted on

[02:43] our social media platforms. But as you see here, we got our nice little tweet here with the hashtags, with the emojis, and just an overall structuring of the tweet we have requested. As you may already know, you can go ahead and edit the original prompt here and so on. But without further ado, let's go ahead and jump over to Banner Bear. All right, when you create your account, you're going to be given a page like this. What you can go from here is going to go over to projects. You're going to go ahead and create a new project and title it however you want. You do my first project, I'm going to go ahead and jump

[03:14] into the course templates that I've been using in the other courses. If you're interested in getting more complex tutorials on these kind of topics, go ahead and check the courses in the description below. I have a whole social media management course that goes over the whole nine yards when it comes to automation, social media management, and using ChatGPT in it. Creating a whole social media schedule based around this. Without further ado, let's go ahead and start and you're going to want to start with creating a template. So it's going to ask you what kind of template you want to create. Custom Pinterest. And these are the different resolutions. Let's say we want to go ahead and get a template that's

[03:46] already being used. So in this context, we're going to go ahead and browse the template library. From here you're going to be able to get the categories. Maybe you have a fitness niche here's, a nice fitness niche, template, all the different templates that can be accessed. The type, blog, banner, certificates, label, Instagram, Story, so on. These are going to be a great starting point and then you can obviously edit them how you want to edit them. So today we're going to dive into creating an Instagram post. Actually, let's go ahead and create a Twitter post.

[04:19] It's going to give you the best resolution. Stats go well here and you can choose your different styles, but we'll go ahead and use the ones that are currently afforded to us. So from here we're going to go ahead and choose this news announcement with photo and from there you can be able to choose a project that you want to add it to. So for me, I'm going to add it to Course templates. And now we have our fully API photo that we're going to be able to manipulate. Now what I mean by that is you can go ahead and hit Edit Template here. And we're going to go ahead and call this template Banner

[04:53] Bear Video. Save that template. And as you'll notice here is Banner Bear has a bunch of layers that you have access to and you can click here and see that they're being highlighted over here. And what essentially what you'll notice is they all have their names for the specific layer. And what we're going to be able to do is we're going to be able to automatically add different data within these layers. So for example, maybe we want a quote here. We're going to be able to use GPT to automatically push a quote here and then push it towards a Twitter feed. And this will make a lot more sense.

[05:24] But for now we have our Banner Bear template. Let's go ahead and jump over to Zapier and start creating that flow that we described before. Now if you're not familiar with Zapier, zapier is an automation software that allows you to essentially automate task. It allows you to integrate data from different platforms and pass that data through. In today's video, we're going to go ahead and create a fully AI automated tweet with AI content associated with that tweet. So to start off, I'm going to use a trigger of Basecamp. What I would suggest you do in the context this is going to be part of a social media schedule.

[05:55] It's going to use Schedule by Zapier and then you can choose every day, every hour, every month, every week. From here you can hit Continue and then you'd be able to choose the time of day. But in the context of this video, since I want to be able to control and manually trigger it, I'm going to go ahead and choose Basecamp which is a CRM. And all I'm essentially doing is setting the trigger that essentially when I make a new to do it will basically cause a flow to occur. This is a great debugging tool to use whatever you want to use as your trigger.

[06:25] This is my trigger when I debug different flows that I'm using for clients or for personal use. So now we've set up our trigger. Let's jump into the real meat and potatoes of this flow which is going to be essentially setting up a fully automated AI tweet with AI content. So to start off, we're going to go ahead and call ChatGPT and just preface real quick. If you want to understand how to integrate ChatGPT or you want to understand how to get started with Zapier, I go ahead and suggest you to check out that video up there. This shows you how to implement AI with automation.

[06:58] Real Quick tutorial gives you an idea of the costs associated and everything that goes along with making the ability to use AI in automation as seen here. So what we're going to do is we're going to hit conversation. Now what's great about this is this is going to allow you to create a memory key, which is going to be very fundamental. You're going to be able to learn a lot with that. So first we're going to do this. I'm going to go ahead and say context, we're going to say this is a business tweet for a business, we're just say for

[07:32] a business called Joe's Coffee. And then we're going to go ahead and ask it to generate a Tweet that is based on and then we're going to do semicolon parentheses. And then from there we may want to add some parameters. We're going to add parameters here. We're going to say generate just one sentence in order to

[08:03] mitigate the possible scenario here where the output is too many characters causing the tweet not be able to execute. So we're going to generate one sentence and from there we can go ahead and choose our model. Now, for more complex, comprehensive situations, you're going to want to choose a higher level model which is going to be GPT Four for more labor associated executions. So summarizing formatting the underlying data. You're going to want to use GPT 3.5. Think of that as your grunt AI, where you're going to want

[08:36] to do tasks fast. It doesn't really need a lot of thinking. For your more complex and intricate AI, you're going to want to use GPT four because for this example, we are generating a tweet. So we're going to want to make the AI as smart as possible in the context of the content we're pursuing with. So here you can change this. As you are a social media manager, you're a helpful assistant. This does help the output a little bit, but it's not significant enough where I wouldn't need to do it right now. Temperatures, obviously you can read it right here, basically zero being higher temperatures, like 0.8 or more

[09:09] random. I would just stick with 1.0. Lower is less random, more fixed, less creativity. Higher, more creativity, but could result in less accurate outputs than when you were originally looking for. So in the memory key, what I like doing here is you can either add random characters, but for this example, we're going to say Tweets and ten. And essentially what the memory key is in this context is the chat that you saw in the beginning of the video. It keeps a record of what is said previously and why that's valuable is that when you're doing stuff like social media

[09:40] management, you want to make sure that what's being said isn't being regurgitated and reused as obviously that will lead to a deterioration of the content that's being consumed by whoever is your follower, right? So from here you might be asking yourself, okay, well, I have a problem. You're generating a tweet, but you're not giving me what it base it off of right now. You can make this a not so smart way of doing it would be essentially base it on coffee or whatever. But what I like to do is I want to make sure that when I make my content, I never have to touch it again.

[10:11] I want to make sure this thing is smart as possible and as automated as possible. That's what we do here at Web Cafe AI. So we're going to add another one here. I'm going to rename this. We're going to say idea creator. We're going to do conversation. And for the Idea creator, we're going to say generate a random topic about coffee. This could be anything. You're in this context, we're dealing with Joe's Coffee. So we're a coffee shop, we're a cafe. Maybe we like random facts, but for now, we're going to do generate a random topic about coffee.

[10:43] We have parameter here and we're going to say, don't use the same topic you have generated before. And in order to ensure that this parameter works, we're going to want to make sure we add a memory key here. We're going to say Idea Coffee as our memory key, any string of characters under 32. And then obviously, because this is a more complex use of GPT, we're going to want to use to model four as we want to make sure that our topics aren't being repeated and it actually understands what it means when they say, don't use

[11:15] the same topic. You can use GPT-3 in this, but end of the day, if you're asking it to do something complex, we want it to have more comprehension of what it's doing. So let's go ahead and let this load. All right, I went ahead and selected it, and one thing I want to point out real quick, which you might notice is this 3114 or Turbo 30. These are basically pushes of updates of the model. They will not be used long term. So when you're choosing a model to choose, always use the ones that's GPT Four, GPT 3.5, Turbo.

[11:46] And as we grow with this AI, obviously we'll get better and better models. But for now, GPT Four, we're going to have to hit continue here and we're going to go ahead and test that action. And then based off this, we're going to be able to get a very concise output. So we're going to say, exploring the Art and science of Aeropress coffee brewing Techniques. So there's our topic. Come down here and we're going to base our tweets on this variable topic. So now, every single time we run this tweet, it's not the same thing of just the same generic topic.

[12:18] So therefore, over time, the content might be the same to the follower. You might unfollow be disengaged. Now we got a variable topic. We're going to be creating a tweet that will change every single time. So let's go ahead and see what this tweet looks like and see if we have to make some edits to the underlying tweet. All right, perfect. So we have our tweet right here. Two things. First thing is sometimes what you'll notice is when you use higher models like GPT Four, the loading time might take a little longer. That's just obviously to do with the fact that we're dealing with more comprehension here.

[12:49] So it's going to want to give you trying to understand more what you're trying to do here. GBD 3.5, as I said before, is for faster types of things like summarizing formatting the data, not necessarily creating the initial idea of what the data should be. So as you'll notice here, we have a little issue. We have an output with quotation marks. And quotation marks never look good in the context of a tweet, might confuse your followers. We're going to format this. We're going to formatter I'm going to rename this. We're just going to rename this as tweet. As with all these videos, you can find the link to the Zap

[13:21] in the description below. If you feel like you've learned anything up to this point, go ahead and make sure you subscribe. Make sure to like liking pushes this content to further people helping us at Webcast AI. So we're going to go ahead and take the formatter tool. We're going to come down here, we're going to do text. And then what I like to do in these contexts is I like to rename them. I'm going to say remove quotation marks. And then the text we are going to be transforming here. We're going to be using replace function. So we're going to input the data of the tweet right here.

[13:51] And then basically we want to find the quotation mark and we want to place it with nothing. And then from there we'll be able to get our tweet with no quotation mark. All right, looking good. So we have created the text version of the tweet moving on from here. And if you are interested in this part of the tour, go ahead and skip over to the Banner Bear part. Give you more in depth here. I just want to give you a full rundown of the utility behind this and why you would use this in the context of Banner Bear. So from here we're going to go ahead and create our Banner Bear image.

[14:22] So we're going to use Banner Bear and we're going to say create image. Continue. We're going to choose our account and for us it's going to be course templates. Great. And from there we're going to choose our template ID. And if you remember, we named it Banner Bear. Let me see what we named it real quick. And from there we named it Banner Bear video. And as you see here, all those layers got added over and

[14:53] this is where you basically enter in the data. So let's go ahead and manipulate, start editing that Banner Bear template. Make it look good, make it look themed to the company we're dealing with here. All right, so first thing we can do here is we can add maybe fixed variables. So for example, maybe we know what our handle is. And it's not a variable in the sense it's going to change every post, but it's going to stay the same. So we're going to do Joe's Coffee here and then from there we're going to go ahead and grab the URL. URL is not going to be variable.

[15:25] Then you're going to add your logo here. So let's go ahead and upload a logo. All right, so we have uploaded the logo here. Now sometimes it will come in a little large like this. You just come down here and you can mitigate the size like this, just dragging like this. So think of this like any type of editor, like canva and so on. This is going to be a great tool for you to use. Basically, think of it like your canva, but your ability to add variables so that every time you push out a video, every time you make a social media post, you don't have to go into canva, redesign the text, redesign everything.

[15:57] This can be all automatic and all done with the magic of Bander Bear. All right, perfect. So we have the logo up here, logo down there. Obviously, edit these how you want to edit them, but in today's context, this is how we're going to edit them. So we're going to do that. And from here we're going to put our brand name, Joe's Coffee. And from there we can go ahead and maybe say Cafe with the best now coffee. Now obviously, as I said before, all this can be changed.

[16:27] So for example, this layer right here called Heading to that could be put with AI. I could put a specific phrasing here if I wanted to. And for today's video, let's go ahead and do that. Let's go ahead and show you two different variables we can put in text wise. And we're going to put in an image variable. So we're going to learn a lot here. So first thing we're going to do here is we're going to rename this layer. We're going to say under Joe's Coffee subtitle. These things can get long here, but just for my sake, we're going to do that. Stagnant, stagnant, stagnant, stagnant.

[17:00] This is going to be our image container. We're going to go ahead and just rename that to choose image. And this will be for our sake, we're going to go ahead and put the caption here. So we're going to put Titan subtitle. We're going to rename this to caption. Obviously, as you'll know, you'll be able to put any AI there if you wanted to. So what do we have? Let's go ahead and change the coloring of this real quick because let's make it all right. So I went ahead and got the color for that.

[17:31] It looks a lot better. Logo is much more matched in there. You can obviously change the coloring here if you wanted to. So maybe we want to make this something like that, so and so forth. But now what you'll need to do here is say save templates and let's go ahead and add an image variable, our caption and our subtitle. I'll see you over at the tab. Perfect. So these data needs to be reloaded, but for now, let's go ahead and add a couple more things here. First thing we're going to want to add here is going to be a Daly prompt.

[18:04] The way we add a Daly prompt is going to be very simple. We're going to go use OpenAI and we're going to say generate image, continue, continue. And then from there we're going to do a size of 1024. And what's great about this is what I implore you to do essentially, is essentially set up a preset of branding style you want for your AI images so it isn't just putting coffee here and then it chooses the style. Now, what do I mean by this? Let's go ahead and jump over to Prompt Base real Quick, which is a platform that sells

[18:35] prompts and you're going to be able to search through and maybe find an AI image style that you really like. All right, so I went ahead and put Prompt Base in the description below as well. This is a marketplace that essentially sells a bunch of different prompts that you can use. So in our context, we're going to be looking at the best dialy prompts. And what I would implore you to do, and you have your little settings over here, go ahead and search through this. Find one that you really like. Maybe you want an iOS app icon concept, clay emojis, 3D product logos. I would implore you to basically find one of these.

[19:06] And these are very affordable. I mean, you're looking at $2 and you can use this real Quick as a reference to see the examples of the photos. And what it does essentially is it gives you a preset amount or preset prompt that you can go ahead and use. And what you'll see later in this video essentially is how you can apply that to your own OpenAI. One thing I want to point out here is webcafe AI does have a portfolio here. We specialize more in selling ChatGPT prompts. So as you'll notice here, you'll be able to find a plethora

[19:39] of specific ChatGPT prompts that you can start using in your business today or in your personal life. For example, a very popular one of ours is going to be the Earn Money Now prompt, which essentially allows you to choose a brand topic or niche, plug it in and get a bunch of ideas of how you can monetize that. Let's go ahead and jump back over to OpenAI and Zapier and figure out how to utilize one of these DALLE prompts. All right, so I purchased one of the DALLE prompts and I'm going to go ahead and use it right here. So all you need to do is paste the underlying DALLE prompt

[20:10] that you got and received. From there you're going to see these square brackets. This is where you're going to enter in essentially what you want being shown in every one of these AI images, right? So in the context of what we're dealing with here, I could make this a variable based specifically on the topic. But for this video alone, we're just going to put coffee here and be satisfied with that. But we get more complex in our other tutorials on how to make that a variable that maybe you want to really make the image based off the title or what's being focused on.

[20:42] But for now, we're going to go ahead and say coffee. This tells you basically how many variations. You're always going to want to choose one and you want to choose the highest resolution. And let's go ahead and say test action. All right? So from the style that I chose, which is kind of like an emoji, we have our coffee cup here. Now obviously there is a plethora of styles, different ways you can approach this. But for our context, there is our emoji. We are dealing with a coffee topic. So we got our coffee emoji from there, we're going to go ahead and make that little subheading. So let's go ahead and start doing that. We're going to do ChatGPT and we're going to go ahead and

[21:14] conversation. And in this context, let's go ahead and use Jet GPT 3.5 as we're going to go ahead and summarize the topic of the article or the topic of the post in three to five words. We're going to say summarize in three to five words, semicolon parentheses. And then we're going to go ahead and say the idea created parentheses again. And then for this context, some random character of words. Just make sure we know and can track this.

[21:44] Because what you'll notice essentially is every time you run it into a chat, the way it outputs it, every time it'll output like that. What I mean by that is, for example, outputs, arrow, press, coffee brewing techniques, some context, maybe it puts it as summarized semicolon and then output and you don't like that. And what I would suggest you to do is instead of maybe messing with this a little bit, just change one of these numbers and look for a different way of seeing how the output could occur. Then what that means for you is that from every time here on out, it will always occur like that.

[22:16] I'm not sure if that made too much sense. If it doesn't, go ahead and let me know in the comments below. But from now on, let's go ahead and take this and plug in our data. So we're going to go ahead and do a couple of things. The first thing we're going to do is we're going to reload this looks like we're doing that. All right, so it looks like the name didn't save on the end, but end of day we're dealing with the same template. So we're dealing with news announcements with photo. We're going to go ahead and say choose Image. What image are we choosing? We're choosing the image URL. Then we come over here under the subtitle.

[22:47] So this is the little header that we just made. Go ahead and choose the assistant response message. And then finally for the title, which should be called Caption perfect. We're going to take the Tweet caption. Obviously, you can formulate anything we'd want a different variable for. In this context, we're going to choose a caption and we're not going to choose a caption with the quotation marks. We're going to choose the formatted caption perfect. And from there, we have officially used Banner Bear to create an AI automated image. And we're going to run this again through basecamp so you

[23:17] can see how this works. If this were, for example, to be used on an actual scheduler, as you'll see. But for now, let's go ahead and see what this image looks like. All right, perfect. So we have our AI image, we have the caption created, and we have the subtitle that was created right down here. Then we have the preexisting branding that we created on our own that are the fixed variables. But now you've created your first AI API image. Really cool stuff.

[23:47] Let's go ahead and create a tweet with this. And then let's go ahead and run it and let's see what a full flow looks like from here. We're going to go ahead and choose create Twitter, not Tweet Twitter. If I can spell do Twitter. We're going to choose create tweet. Continue. And I'm going to go ahead and do this on our main account so we can really see it full fledged out. And then I'll delete it, obviously. So anyone following is going to be a little confused.

[24:18] You want to know when new articles come out and anything related to this kind of stuff, go ahead and make sure you follow us on Twitter. Then we're going to go ahead and put in the formatted caption. Now what's cool here is you can also add other stuff here like follow us. And this will be a fixed variable. So every time there's a tweet, it'll say follow us. From this context, I'm just going to do the caption, and then we're going to choose the image that was just created right down here. Always go JPEG, less megabytes dealing with better and better. And then from there you can hit Continue test

[24:50] action, and then it will shoot off. But what I want to do, and I want to show you guys is I'm going to go ahead and launch this as. If let's say it was Tuesday at 08:00 A.m. And it was going to go, we can go ahead and see what it looks like fully automated. So we're going to skip this test and I'm going to go ahead and publish this. Publish and turn on and then let's go ahead and jump over to base camp. Now over in basecamp, this is one of our sandbox accounts. All I need to do to trigger the flow is I'm just going to type in literally anything. Going to hit Enter and let's jump back over to Zapier.

[25:21] As you saw before, we can go ahead and click History here. And then in History we see that it's currently playing. So you can click this and you'll be able to track the steps as they go. You can hit data out. And if you want more information on this, I implore you to go ahead and take our mastering business, which at UBT we go over a lot more complex stuff like this or get a little taste of it with our free course which goes over stuff such as monitoring and optimizing flows like this. Because as you'll notice, tokens are expended.

[25:52] But we're talking about fractions of pennies here. There's different ways to understand scaling wise, how much one specific flow cost. But from here you see it's playing here. Let's go over to and as you see here, it is successful. Let's jump over to Twitter and see what it looks like. All right, so we are over at Twitter and as you see here, we got a pretty cool looking tweet here. So we got our Discover the fascinating influence of Grind Size on your coffees flavor profile at Joe's Coffees. And let us guide you in mastering a perfect grind for a

[26:22] truly personalized taste. Nice hashtags. Then we have our nice tweet. I mean, look how good that AI image looks. I mean, that just looks amazing. You got our tweet again. And then we have grind size effects flavor. Everything's looking good. Everything's looking great. There you go. There is your fully AI automated tweet. And now let's go ahead and jump back over to Zapier. All right, so I'm going to go ahead and share this Zap in the description below. If this was really interesting to you and you want to learn more, go ahead and check our courses below.

[26:52] We have a social media management course that goes over a lot more complex topics than this and shows you how to make a fully AI automated schedule based off the content you want to create. Plus using user created content that maybe you want to put in and have it edited automatically. Really cool stuff. So all the resources can be found in the description below. If you felt like you learned something, go ahead and make sure you like the video. Go make sure make you subscribe. We do daily videos here showing how to use AI automation in your personal life and your business life.

[27:23] Now go ahead and click that video there if you want to look how to use AI automation, specifically video and specifically Cynthia. So it's going to be a great tool for you to use for scaling when you need to do voiceovers or use actually AI avatars cool looking video there. Without further ado, I'll see you in the next video.