Create AI Videos with VEO3 on Autopilot! (Make + n8n Templates Inside)

So this is an AI video about nothing. It's about nothing. Who would watch a video about nothing? When I made that comment, I wasn't saying all kobolds are bad and stinky, just the ones from Goblin Town. And it's true. It's true. Guys, what if your next YouTube video, TikTok Short, or viral ad can write itself, produce itself, and publish itself automatically without you having to lift a finger. This is the automation that I'm going to show you how to set up today. And best of all, I'm going to show you how to set this up on Make.com as well as N8N, which are the most popular platforms these days for building automations as well as agents. Now, if you're new to my channel, hi, my name is Helena, and on this channel you will learn everything about AI automations, So please, please help this channel out by giving it a like and subscribe. And on top of that, I also have a free course that will go deeper into the topics that we talk about here today. So if you want a deeper dive into AI agents or automation, sign up for my free course below. Now, the tool that we're going to be using to build these AI videos today is Veo 3. It's the latest AI video generator from Google. It has completely gone viral with just one text describing what kind of video you want to create. Can now create CGI-level videos and it's absolutely blowing everyone's mind. Now, this tool is quite expensive though, it's $250 per month. And as AI video creation gets better and better, I highly suspect that in 6 to 12 months from now, this price is probably going to be halved, if not like just 10 to 30% of what it is today. But today, as of right now, it is still quite pricey, but If you use the API of VO3, you only pay for how much video you generate. And at the moment, if you don't use audio, it's $0.50 per second or $0.75 per second. No, not per minute, per second. Yes, I know it's pricey, but if you do get a human to create similar levels of video using CGI and animated characters, it's going to be at least $20 to $30 per second. So using AI, you will still be saving money. But I do understand that compared to other AI video generation APIs, this is definitely on the higher end. Now you can create the entire automation that I'm showing you here today though with any other AI video generating API as well. You just swap out the VO3 for the other APIs, then you'll all be set to go as well. With all of that being said, let's jump into my screen and let's make this automation now. So let's start with make.com and let's start making making this video generating automation on Make here. What I do like about Make is that it already natively integrates to thousands of different tools, whereas N8N doesn't do this. So the trigger that we want here is ChatGPT, actually, because this is the first step of this automation. We want the prompt to be written first, and then we can generate the video, and then we can get the video and then post the video. So that is why ChatGPT would be the first trigger. Now, there is another way that you can set up this automation. You could have the trigger be a new row being added to a Google Sheet or an Airtable. So in that case, a whole automation will be triggered whenever you add a new idea to the form. It's not purely automated at that point, but if you want this whole process to be fully automated, the trigger here would be the OpenAI module. Now, once you get into your OpenAI module, and if this is your first time connecting to your OpenAI module, you will need to get an OpenAI API key. To get that, we're going to go over to platform.openai.com/api-key like I have here. And then we're going to go up here at the top and create a new key. We can give this key a name and click on Create the Secret Key. Then we're going to copy this API key here, go back to our Make automation, click Add. Paste in this key here and then press Save. I have already done that, so I don't need to do it again. And then for the model here, we can select the OpenAI model that we want to use. Now, the reason that there are so many different models is because they have different amounts of capabilities. So generally, the more pricey the model, the better the output that you're going to get. And we can check the pricing for these different models by going to openai.com/api/pricing and we can see all the pricing here. So for what we need to do here, we don't need the model to be the most powerful one because we just want a prompt back. So that's a pretty easy generation. And for this case, I'm going to use GPT-4o as my model. And then now it's time for me to write my prompt. I just said expand on messages. The role would be user, and this is where I put in my prompt. And now a little cheat or a little hack here. You don't have to write the prompt. We can actually go into ChatGPT and have it write the prompt for us. So to help ChatGPT write my prompt here, what I did was I actually just added some images of the viral orange cat and the babies that were AI generated that was absolutely going viral. Like, it's getting like hundreds of thousands of likes on Instagram and TikTok. So I just added some screenshots of them here just so that ChatGPT has an idea of what I'm looking for. All right, I have my prompt here ready to go. So I put in, I am looking to create AI-generated videos like the screenshots that you see here. I'm creating on Veel 3 and I need your help writing the prompt. The prompt should be catchy and 2 to 3 sentences only. This prompt will go into my Make automation and create a new scene daily. And then we're just going to wait a second for AI to get back to us with the results. All right, so we can see here it's thought about it and then it's generated here the prompt that we can use. I'm just going to copy this and go back into our Make Automation, paste that in. The only thing I would change is from 7 seconds to 8 seconds. And then we can't put the curly brackets behind this, so I'm just going to Put in the square brackets and put in insert scene here. All right, now I have put in my prompt. I just added in an extra bit here because when I was testing with the VO3 automation, I actually found that whenever there were the punctuation marks added in, it actually breaks the entire automation. So I told ChatGPT to not include these punctuations as well as keeping the prompt to just 3 sentences long. Everything else can stay the same, and I'm just going to save this here. And then before I move on to the next step, what I usually like to do is actually just to give everything a test to make sure that everything is running properly. All right, after ChatGPT has finished generating, we can actually click on the little one and expand with the magnifying glass and see what the result here is. And I do like the results, so it is working properly. And now we can move on to the next step. So the next step is to use Veo 3 to generate this video now that we have this prompt. And so to do that, I'm going to click on the plus button and the API consolidator that we're going to use for Velo3 is something called Fal.ai. Fal.ai is an API consolidator. So basically by having one Fal.ai API key, you'll be able to use hundreds of AI tools off of their platform. So instead of you having to get a Midjourney API, a Flux API, a Velo3 API, you can just get this one API key and then you'll be able to use all of those tools off of their platform. It's quite good and I do really like it. And so once we get into file.ai, we can see all of the different actions that we can take. And what we want is make an API call because turning text to video using VO3 is not a native action that they have here. So once we select API call, now if this is your first time setting up your file.ai account, first of all, you do need to fund your account so that you can start having some credit inside of your account. What we want to do is go up here, go to our account, and then go over to API keys. And this is where we can create a new key, call it something that we want. Now that we have our key, we can copy this key, go back to our Make automation, press Add, and this is where we paste in our key. And then now we're ready to assemble the rest of the sections here. Now for the URL, this is something that messed me up for a good half an hour. You don't have to put in the full URL. As you can see here under the subtext, the first part of the API URL that you're trying to call is actually filled in for you already. So you just have to put in the last part of the URL. And to search for what you want these URLs to be, what you want to do is go up here and go to search and type in the specific AI tool that you want to connect to, and in our case is Veo 3. So when I click on Veo 3 here, I can see that the ending is just veo3, and that's all we need to put here, right? That's a cool thing. You don't have to put the entire URL, just the last part after file-ai. And the method is going to be POST because we want File AI to create the video. For the header section, we just have one. So we want the key to be content-type and the value to be application/json. And now we get into the body section. And if we go back and look at the instructions from Flow AI, what we can do is go to the playground and then we can see the different formats that we can get for the inputs. We want to go to JSON. And we can see the exact body that we input into our API call. So I'm going to copy and paste this over. Now, we do have to make a couple of modifications here. This is just a sample prompt that they have. So I'm going to delete this out because we want our prompt to be the result that just came in from the previous ChatGPT module. And also we want this to be a YouTube Short, so we have to change the aspect ratio from 16:9 to 9:16. Now, one pro tip here is don't add in all of these variables as a query string because that's what I did. And again, that was something that took me half an hour to figure out what I was doing wrong. So once we do this, we are all set on the cloud.ai module. And now what happens is whenever this prompt goes into Flau.ai, it will then be generating this video in the background. But it does take quite a bit of time to generate because these videos are pretty good quality. So from my experience, it takes anywhere from 5 to 15 minutes to generate. So what I like to do here is actually just let this whole automation wait until this entire video has finished generating. So to do that, I'm going to create a new module and I'm going to look for sleep. Another way that you can find sleep is you go down here to the purple button where it says tools, and you can search here until you find sleep. So that's the 2 ways you can find the sleep module. And then the maximum amount of time you can allow this automation to go to sleep for is 300 seconds or 5 minutes, which means that if it takes this automation like 15 minutes to generate, I will need at least 3 of these sleep modules. I'm going to do 4 just in case because I don't want to be in a situation where this whole automation is still running and it can't proceed to the next step just because this video has not finished generating yet. After we do this, the next thing we're going to do is actually just set up the schedule. For this entire automation to run. So to do that, I actually go down here to the bottom and you see where it says every 15 minutes, right? What I can actually do is adjust the schedule setting and I can adjust how often I want the whole AI system to create a new video. It could be every day, it could be specific days of the week, it could be days of the month, or specific days that you want. In this case, let's say I only want a new video every Tuesday. I could select Tuesday there, and then I can select the time that I want this video to be made. Let's suppose it's 7:57 every morning, and I would just take whatever time I want that video to be posted and then add another 20 minutes to do that, knowing that the video will take 20 minutes to create as we're waiting for 20 minutes here. All right, so once that is done, once this video is finished, what we need to do then is to retrieve this video. And to do that, we are going to open up another Flow.ai module here. And instead of making an API call, this time we actually want to get a response. And we can see we had to put in the request ID here. And because we haven't ran this entire automation yet, We can see here that I actually don't have the request ID. So if you just want to test out this entire automation, you can actually press Run Once here and it will run the entire automation. After you have ran the automation, you can also go into your Flaw.ai account and go over to the Request tab and you'll be able to see all of the videos that you have generated previously. So this was one that I generated of the cat and the baby and you can see How cute is that, right? So, and this was all AI, guys. Like, that's, you know, just the thing that continues to blow my mind. This is another one that I generated here. So you can see that, like, in terms of quality, this is definitely top-notch for a purely AI-generated video, right? Once you have ran this automation once and you go to the Flaw.ai module here and under Request ID, you'll be able to see more options here, and what we want is the Request ID. And once we get that, basically, it will give us now the MP4 URL to the video that we have just generated. And right now, after this point, we have 2 different options. So the first option is if you want to review this video before it is posted, What you can do is to get this MP4 link to be appended to a Google Sheet, let's say, or send you an email. If you want it to be sent to your Google Sheet, we're going to look for the Google Sheet module and we're going to select Add a Row. I created this sheet earlier and basically it just has 2 columns, one for the video idea and the one for the video output link. And if we go back to our Make automation here, We're going to look for this spreadsheet, and we can simply just fill out the different fields that are involved here. So the video idea will be the prompt that we put in, and it will be the output of the ChatGPT module. And then the video output here, if we expand on get a response and then expand on images, we will be able to get the URL of this video. Okay, so that's the MP4 video that Flow was able to create for us. Now, this is what we would do on Make.com if we want some sort of like human in the loop, some sort of human oversight. If you don't want to do this, another option would be that you just automatically post on all your socials from here. So to do that, if we wanted to post on Instagram, we just look for the Instagram module. And then we select Create a Reel Post. From here, we make sure we're logged into our Instagram account, and we just need to put the URL of the video that we want to post. And again, this will come from the Flow AI Get a Response module. And then we expand on the images, we select the URL. We can add in a caption here. If you want the caption to be the same every single time, you can just type it here. Or another thing you can do, is actually add in a ChatGPT module right here for it to write the caption of your Instagram post and then simply select the result from that ChatGPT module and put it over here. This is not a mandatory field, so you don't even need it here as well. If you want it to be posted on more of your accounts, you could simply add a router. So a router is when you click on the green Flow control at the bottom, and you want to select a router and link this to the Flow AI instead. And then what you can do is link this to all of the accounts that you want to be posted on. So this could be your Instagram account. It could also be your Facebook page if you want to do that. Select the Facebook module and then go into Facebook pages. And then for the actions inside your Facebook module, make sure you select upload a video. And then the same process as the Instagram module that you saw earlier, simply put in the URL of the video here and we can put the name as just a prompt. It doesn't really matter. And then that's it. And you can put in a little description here if you wanted to. So if you do want a different caption for each of the platforms, Simply unlink this and add in another ChatGPT module. That will work as well. You can also automatically link this to your YouTube account. Just add in a YouTube module here and for the video to be uploaded automatically, same process as before. So this is how you would set up this automation on Make.com. Now let's go over to N8N and I'll show you the differences here. This is the same automation. That I recreated instead of N8N. So the trigger that I set up here is a little different. Now, I told you, like, whenever it comes to these automations, there are multiple ways that you can make this. So for this automation, I just wanted to show you another possibility. For this, it's whenever there is a new row being added to this Google Sheet here, this whole automation will be triggered. And then what's happening here is that We are writing our prompt same as before. What I did here was just dragged and dropped an OpenAI module here. So one of the drawbacks of using NaN versus Make or Zapier is that they don't have as many natively integrated blocks with other softwares available. It doesn't mean that you can't use any of those softwares on their platform. It just means that you have to write the API integrations yourself. You can see here their list is quite limited. There's only a couple hundred natively integrated actions inside of their platform, and Flau.ai unfortunately is not one. So what we need to do is actually look for an HTTP request and then fill out the URL, the header, and the body ourselves. I already did this here, so I'm going to delete this and get rid of this one here and link it over here again. And basically here you had to put in the the entire Flow API endpoint for it to run. And then you had to put in the other header parameters and the JSON body parameters here as well. Once you do that, a different way that I have set this up is for the automation to wait 5 minutes and then it's going to check the status of the video and see if the video is completed or not. If it is completed, we're going to go down to the next step, and if it's not completed, it's going to go back and wait another 5 minutes, and it's going to check the status again to see if it is completed. So basically, this is a little different from the way that we set up here where we're just waiting 20 minutes and continuing with the automations. I'm just showing you different ways that you can have this entire automation set up. And now, after the status is completed, What we will need to do is to get the MP4 URL. Now, instead of having these nice blocks that we have natively set up in Make.com, we have to make all of these custom HTTP calls instead of N8N. And then the last part here is to just update the sheet. So when this MP4 video is finished, then it's going to populate that into the Google Sheet that we see here. So you may be wondering, okay, N8N looks like it's a lot harder to use, so why would anyone consider this? So the main functionalities of N8N is that it has human in the loop, and it also has all of these advanced AI agent modules, which can be really useful depending on your use case. For this particular use case, we didn't really need any of these advanced AI nodes. Where I see this coming into handy is we can add a human in the loop. And let's say it sends you a Slack message or Telegram message. So where you can like approve or disapprove of the outputs that came in. Once the video has been approved, it's for it to be uploaded onto your Instagram, Facebook, and YouTube account. So it saves you that little bit of extra step. Than what you have to do on Make.com of like approving the video and then manually uploading it onto the appropriate social accounts. So those are the main differences between Make.com and n8n. I hope you enjoyed this tutorial. I will also include a link to this automation below as well so that you can just import it into your account and give it a try. Thank you so much for watching this tutorial. I hope you got a lot out of it. I just absolutely love VO3. It's such a game changer and I I can't wait to see where this industry goes in the next few months. If you haven't done so already, make sure you sign up for my free AI and Automations course as well, and I will see you in my next tutorial.