WEBVTT 00:00.240 --> 00:05.440 Recently, I came across this YouTube channel that gets millions and millions of views, and one of 00:05.480 --> 00:08.560 their videos made 407 million views. 00:09.000 --> 00:10.760 Really, this is the real number. 00:11.000 --> 00:13.080 And let's watch what we have inside. 00:14.120 --> 00:19.080 So even we don't have any background music like this is the cat trying to trap a mice. 00:19.200 --> 00:20.880 So this is pretty cool. 00:20.880 --> 00:25.640 So I thought I would create an automation that can generate this type of content. 00:25.800 --> 00:27.760 And I came with something like this. 00:27.760 --> 00:33.240 So the simple automation that doesn't contain a lot of steps however it generates for me this type of 00:33.240 --> 00:33.920 videos. 00:34.520 --> 00:37.080 So this is the video I created with this automation. 00:37.200 --> 00:40.640 And the cost of generating this video is approximately $0.50. 00:40.640 --> 00:41.840 And it's really high quality. 00:41.880 --> 00:44.000 We've got the mice, we've got the cat. 00:44.000 --> 00:45.920 And again it's highly professional. 00:46.520 --> 00:51.720 So in this material I will guide you through step by step how to build this entire automation to generate 00:51.720 --> 00:56.600 these videos so you can publish them on any social media you want, let's say TikTok YouTube shorts 00:56.600 --> 01:00.160 and get millions, thousands or even hundreds of views. 01:00.280 --> 01:05.910 So again, this channel gets thousands of views and the most popular videos get millions of views. 01:06.230 --> 01:12.110 If you watched other automations in this course, such as click 2.1 plus Naarden or even Vo three Content 01:12.110 --> 01:18.030 Machine, you can notice that the steps inside these automations and this automation I will show you 01:18.030 --> 01:20.430 how to build in this video are pretty similar. 01:20.630 --> 01:23.550 From this simple fact, we won't build the entire automation. 01:23.550 --> 01:27.390 In this video, I will guide you through the entire process of how to build it. 01:27.430 --> 01:33.550 However, again, if you would like to go and learn how to, for example, connect your Naarden with 01:33.550 --> 01:34.990 Cloudinary open AI. 01:35.390 --> 01:38.670 And here we've got full AI to generate the video. 01:39.230 --> 01:45.710 You can go ahead and watch this material about the click 2.1 plus Naarden or even Vo three content machine 01:45.710 --> 01:47.790 because again, we've got pretty similar steps. 01:47.790 --> 01:53.030 For an example, in the Vo three content machine we've got full AI, and here also we've got the HTTP 01:53.030 --> 01:55.310 request for full AI right there. 01:55.310 --> 01:57.470 Also we've got let me see. 01:57.830 --> 02:02.470 Um, even here we've got the method of how you can publish the content. 02:02.470 --> 02:08.940 So for an example if you wish instead, or even for the further steps, you can provide the data and 02:08.940 --> 02:11.700 then publish the content on TikTok or YouTube shorts. 02:11.940 --> 02:15.660 So again, in this video I will guide you through the entire project. 02:15.700 --> 02:22.300 However, we won't build it because it's kind of wasting our time to repeat setting up the same notes 02:22.340 --> 02:25.100 over and over again and through different automations. 02:25.180 --> 02:29.900 That's why, again, I will guide you through each step and it will be very clear. 02:30.020 --> 02:32.060 So with all that being said, let's dive in. 02:32.340 --> 02:35.380 The first step we've got in this animation is schedule trigger. 02:35.580 --> 02:40.940 So actually I like to use it when it comes to the content generating some articles I don't know, maybe 02:40.940 --> 02:43.900 YouTube videos or even any content. 02:44.100 --> 02:46.740 So to begin with the first step is our trigger. 02:46.860 --> 02:51.700 So here I like to use the specific trigger when it comes to publishing, like creating any content, 02:51.700 --> 02:55.980 you know, maybe some images, videos, posts on LinkedIn and so on. 02:56.140 --> 02:59.900 So I run it like every day or every second day. 03:00.060 --> 03:05.580 So you can choose even like let me show you, you can click on Add Note and add another trigger. 03:05.580 --> 03:09.930 You can just add a other triggers such as trigger manually and let's say on a form. 03:09.930 --> 03:11.090 So it's your choice. 03:11.090 --> 03:12.970 My preference is schedule trigger. 03:13.610 --> 03:19.810 Um, right after our trigger, we've got our a agent that creates for us the prompt. 03:19.810 --> 03:24.410 So let's preview what we have and the prompt we've got create a prompt to use for image generation. 03:24.410 --> 03:29.890 And by the way remember in the resources of this material you've got the JSON file to download for this 03:29.890 --> 03:33.330 entire project with the instructions with the system messages. 03:33.330 --> 03:36.370 With all of the steps set up, you need to only set up your credentials. 03:36.410 --> 03:36.810 Okay. 03:37.210 --> 03:39.410 However, um, let's go back to the point. 03:39.410 --> 03:44.210 So in the prompt we've got create a prompt for um to use for image generation. 03:44.490 --> 03:49.970 And in the system message we've got the purpose capabilities interaction style and example output prompt. 03:50.130 --> 03:57.410 So we exactly specify the theme of the image because later will turn this image into the video. 03:57.410 --> 04:02.730 So the theme of this image is a cat as you can notice like here, every descriptive scenario featuring 04:02.770 --> 04:07.530 a tabby cat sitting and setting a clever trap for a mouse with cheese. 04:07.570 --> 04:11.640 Natural cozy home settings described in the warm, realistic photographic style. 04:11.640 --> 04:15.560 So the point is we are just describing the system message for this specific case. 04:15.560 --> 04:19.360 So these specific videos, because they get millions of views. 04:19.400 --> 04:25.360 However still remember with this automation you can customize the system messages all of the steps for 04:25.360 --> 04:26.480 a specific purpose. 04:26.520 --> 04:30.280 It's not that hard at the end if you have your automation created. 04:30.520 --> 04:37.280 Um, then just setting up your system message like instructions for different purpose is the smallest 04:37.280 --> 04:37.920 problem ever. 04:37.960 --> 04:40.680 Okay, um, because you've got the entire system. 04:41.320 --> 04:43.680 So let's go ahead to ChatGPT and do this stuff. 04:43.920 --> 04:49.120 And for the capabilities design unique image prompt with precise character actions and setting details 04:49.120 --> 04:55.800 include props like kitchen tools or common household items for realism, so I don't want to read it 04:55.800 --> 04:56.120 all. 04:56.160 --> 05:01.080 However, you know this is the entire concept interaction, style, writing clear, rich, descriptive 05:01.080 --> 05:01.440 language. 05:01.440 --> 05:05.880 Avoid headings or instructions in the output, just the image prompt itself and example output. 05:05.920 --> 05:07.320 So we've got our prompt. 05:07.640 --> 05:09.720 And this is pretty everything for this agent. 05:09.720 --> 05:13.750 So again we've got the prompt system message and also the chat model. 05:13.790 --> 05:16.870 Obviously we need to have the chat model to run a agent. 05:17.150 --> 05:19.990 And afterwards we've got the we've got the prompt. 05:19.990 --> 05:21.110 So we've got an output. 05:21.350 --> 05:23.630 Also you can use output parser. 05:23.670 --> 05:27.270 So let's say you specify the name of the variable which is the prompt. 05:27.550 --> 05:32.870 Um and yeah you provide only one um specific variable to return. 05:33.070 --> 05:35.950 However um it's not needed for that case okay. 05:35.990 --> 05:37.830 It's pretty simple straightforward. 05:37.830 --> 05:38.790 So we've got clever. 05:38.790 --> 05:45.590 Tabby cat crouches quietly on a sunlit kitchen floor as green eyes locked on a small mouse near a piece 05:45.590 --> 05:49.070 of golden cheese perched on a wire mesh trap. 05:49.350 --> 05:49.990 Perfect. 05:50.230 --> 05:54.230 So this is our, um, create prompt, um, part of this automation. 05:54.510 --> 05:56.390 Next, what we do, we generate image. 05:56.390 --> 05:58.430 And we've done it many times in the course. 05:58.430 --> 06:03.230 Really many times in the, in the course um, by using open AI image model. 06:03.270 --> 06:03.790 And why? 06:03.830 --> 06:06.590 Because it's the best for now for generating images. 06:06.630 --> 06:07.110 Okay. 06:07.510 --> 06:08.510 So what do we do? 06:08.550 --> 06:09.830 We provide an URL. 06:10.110 --> 06:14.580 Um, and again we set it up many times and other automations in this course? 06:14.860 --> 06:15.740 Um, so yeah. 06:15.900 --> 06:21.940 So we've got our URL where we specify we want to generate image here authentication I use predefined 06:21.940 --> 06:26.300 credential type OpenAI OpenAI account and send body JSON body. 06:26.300 --> 06:28.340 So inside what we have we've got the model. 06:28.340 --> 06:30.900 We've got the prompt quality and size. 06:31.300 --> 06:35.660 So for this type of the video we want to specify the size which is vertical. 06:35.660 --> 06:36.980 That's why we've got the size. 06:37.620 --> 06:38.220 Um okay. 06:38.940 --> 06:39.780 This is the step. 06:39.780 --> 06:40.380 We run it. 06:40.420 --> 06:42.740 We get the b64 file. 06:43.380 --> 06:48.860 And then in some shape of form we need to convert it to base64, um, to the file. 06:48.860 --> 06:49.500 I'm sorry. 06:49.780 --> 06:51.500 So we use convert to file. 06:51.500 --> 06:54.220 And inside we just provide this specific variable. 06:54.220 --> 06:55.860 So we just drag and drop it here. 06:55.860 --> 06:58.540 And afterwards we've got the file which is called data. 06:58.900 --> 07:03.340 Um and we can view it what we have we've got a mice perfect and a cat. 07:03.860 --> 07:04.540 That's cool. 07:04.540 --> 07:09.980 So we'll be animating that turning that from image to video with really nice model which is minimax. 07:10.580 --> 07:14.610 But uh, for now, um, in some way. 07:14.770 --> 07:19.410 We need to take this file and upload it to the hosting platform. 07:19.450 --> 07:19.770 Okay. 07:19.810 --> 07:26.130 Because afterwards we want to just put it into into our model of the full AI, which is minimax. 07:26.330 --> 07:27.570 So we use Cloudinary. 07:27.730 --> 07:32.410 Um, also we've covered that in the course I think um, a few times. 07:32.530 --> 07:34.770 So this is the platform that allows us to do this stuff. 07:34.890 --> 07:40.730 So here inside you need to provide there in this link you need to provide your cloud name. 07:41.290 --> 07:48.890 Um, so pay in mind if you copy my, um, JSON template just inside here, provide your cloud name. 07:49.170 --> 07:52.170 Um, I showed it many times in this course how to do this. 07:52.890 --> 07:56.530 So you can refer to, for example, these videos if you wish. 07:57.130 --> 08:02.850 Um, so you provide your cloud name and in the send body you provide form data parameter type which 08:02.850 --> 08:07.370 is an a then binary file name which is file and and field name which is data. 08:07.410 --> 08:10.930 Because obviously here we've got a data for the parameter type. 08:10.930 --> 08:13.250 We've got the form data upload preset. 08:13.250 --> 08:16.250 And here it's important you need to provide your preset name. 08:16.250 --> 08:17.210 That's perfect. 08:17.250 --> 08:18.810 And just go over to Cloudinary. 08:18.810 --> 08:21.090 And here you will have your preset name. 08:21.770 --> 08:26.090 Uh, maybe let me very shortly show you how it works. 08:26.130 --> 08:26.930 Cloudinary. 08:26.930 --> 08:28.690 So it will take only a few seconds. 08:29.050 --> 08:30.170 Um, let me sign in. 08:31.010 --> 08:34.010 So when it comes to the cloud name, you can copy that from home. 08:34.210 --> 08:36.330 Um, right here at the top. 08:36.330 --> 08:38.250 And you need to provide it in the URL. 08:38.410 --> 08:44.850 And for the preset, you need to click on the settings and then on upload and here at your preset. 08:45.050 --> 08:48.570 So in general you need to copy the name of it and provide it. 08:48.730 --> 08:50.050 Uh right here. 08:50.090 --> 08:50.730 Perfect. 08:51.290 --> 08:52.570 The next step what we have. 08:52.570 --> 08:55.650 So we've got two parts create prompt generate image. 08:55.850 --> 08:59.850 The next step what we do, we actually turn the image into the video. 09:00.130 --> 09:03.490 And here we've got a specific link which is specified for this model. 09:03.490 --> 09:07.810 So minimax Halo zero two Pro image to video. 09:08.010 --> 09:12.010 And here there is something important because let me show you. 09:12.290 --> 09:15.810 We go over to fly and we've got our model which is this one. 09:16.170 --> 09:17.920 So the cost of generating it. 09:17.960 --> 09:21.920 It's approximately $0.50, as you can notice for the six seconds. 09:21.920 --> 09:22.400 Video. 09:22.760 --> 09:28.400 However, if you want to use something cheaper, you've got the model which is image to video standard. 09:29.080 --> 09:30.880 Um, so here it's cheaper. 09:30.920 --> 09:37.440 Also, if you go ahead to API, you pick the table of the content, which is http URL. 09:37.920 --> 09:43.360 Then you go ahead here and you've got the URL you can use for this specific model. 09:43.360 --> 09:46.360 So only probably you need to change from pro to standard. 09:46.360 --> 09:47.040 Let me see. 09:47.480 --> 09:49.320 Um let's see. 09:49.360 --> 09:49.560 Yeah. 09:49.600 --> 09:51.240 You need to change here to standard. 09:51.240 --> 09:53.200 And then you're using standard version. 09:53.560 --> 09:55.960 And the advantage of it. 09:56.160 --> 09:57.880 Um, we go over here. 09:58.000 --> 10:04.680 The advantage of it is that we are we are paying only, um, twice as less, I can say. 10:04.680 --> 10:04.960 So. 10:05.000 --> 10:07.200 $0.27 for six seconds. 10:07.200 --> 10:07.720 Video. 10:08.080 --> 10:10.600 Um, of course, the quality will be just worse. 10:10.600 --> 10:12.120 However, it's not bad, as you can see right there. 10:12.160 --> 10:12.600 Okay. 10:12.800 --> 10:13.800 So we can just use it. 10:13.800 --> 10:18.840 This is the disclaimer if you want to just save money on this automation okay. 10:19.710 --> 10:21.390 So in the URL we've got it. 10:21.430 --> 10:25.550 We've got our model which is generic type header authentication full. 10:25.550 --> 10:32.830 I um so I connected it by providing authorization here, probably the key and the API key as I remember. 10:32.830 --> 10:39.470 Well um, you've got also that in other videos you can go over here and see how you can connect, um, 10:39.630 --> 10:43.870 your full I with an ID, um, and the JSON. 10:43.910 --> 10:47.790 Then what we do, we send the body, we've got the JSON using JSON. 10:48.150 --> 10:50.150 And here we've got our JSON code. 10:50.150 --> 10:57.430 So we provide a URL, JSON URL and aspect ratio which is nine by 16 and a prompt. 10:57.430 --> 10:59.590 So in the prompt we've got a static camera. 10:59.590 --> 11:03.110 We don't have the prompt for for this specific circumstance. 11:03.350 --> 11:06.790 Um, because we've, we had the prompt for only the image. 11:07.190 --> 11:10.750 So we just specify okay, make the camera static. 11:10.750 --> 11:16.430 Because in these videos, uh, you know, like everything is static, cat is not doing like it's not 11:16.430 --> 11:19.870 performing a lot of activities, um, and so on. 11:19.870 --> 11:20.950 And the camera is static. 11:21.060 --> 11:21.340 It's not. 11:21.380 --> 11:22.180 It's not moving. 11:22.460 --> 11:23.140 Perfect. 11:23.940 --> 11:25.300 So it's pretty simple. 11:25.980 --> 11:26.740 Then what we do. 11:26.780 --> 11:27.940 We wait ten minutes. 11:27.940 --> 11:31.380 Because mostly these videos take ten minutes to create. 11:31.500 --> 11:39.620 If I go here to Pro model to requests here, the video took, um, 500 seconds. 11:39.660 --> 11:40.300 Almost. 11:40.900 --> 11:44.100 Um, and there 503 seconds. 11:44.100 --> 11:44.580 Okay. 11:44.620 --> 11:46.380 So we just wait ten minutes. 11:47.260 --> 11:52.060 Um, then what we do, we get the video and we provide the specific URL. 11:52.100 --> 11:55.620 So what we specify, we want to connect with minimax requests. 11:55.780 --> 11:58.420 So this specific tab um this one. 11:58.700 --> 12:06.260 And here we just provide, uh, the request ID the request ID um let me show you this is this ID. 12:06.260 --> 12:09.380 So we just specify we want to take this ID and this is the variable. 12:09.380 --> 12:12.380 So every time it changes and we check the status. 12:12.420 --> 12:12.900 Okay. 12:12.940 --> 12:14.420 So it's pretty simple. 12:14.580 --> 12:18.180 And for that case we only provide our credentials methodget. 12:18.180 --> 12:19.540 So it's very important. 12:19.700 --> 12:20.660 And get video. 12:20.700 --> 12:26.050 You've got the method get and post actually an image to video you've got post. 12:26.050 --> 12:29.250 Similarly with creating image, you've got the post here. 12:29.450 --> 12:31.210 Um, yeah, we don't have the get. 12:31.610 --> 12:36.730 Um, but again we've got method get URL credential and that's pretty everything. 12:37.130 --> 12:40.650 And at the end we've got our Google Sheets where we save the data. 12:40.650 --> 12:45.570 So I provided image prompt here, image URL and video URL. 12:45.850 --> 12:48.930 Also you can find it in the resources of this material. 12:48.930 --> 12:52.810 So the specific Google Sheets I need to say it's pretty straightforward. 12:52.810 --> 12:54.530 So we've got only three columns. 12:54.530 --> 12:58.090 The first one is image prompt image url video URL. 12:58.130 --> 13:01.650 So we just just save our data for further tracking. 13:01.970 --> 13:05.730 Um, to just have the access to all of the videos, images and the prompts. 13:06.170 --> 13:08.290 And for this case, this is everything. 13:08.290 --> 13:11.530 I think these videos are highly professional. 13:11.530 --> 13:15.250 We can generate really nice videos with this specific automation. 13:15.250 --> 13:17.370 It's not that hard, as you may notice. 13:17.410 --> 13:22.250 Actually, we've created many much complicated automations in this course. 13:22.250 --> 13:27.320 I mean, much more so in general yeah this is pretty straightforward automation. 13:27.680 --> 13:29.800 And we've got a lot of HTTP requests. 13:29.800 --> 13:33.600 However from this course for sure you know how to use them properly. 13:33.600 --> 13:34.560 Without a doubt. 13:34.800 --> 13:40.600 We've got one agent scheduled trigger weight node convert to file save information to Google Sheets. 13:40.600 --> 13:46.200 So yeah by this way we generated the system that produces for you this type of the videos. 13:46.560 --> 13:53.760 And also remember if you wish you can go over for example to this video and check how you can, um, 13:53.800 --> 14:00.320 actually publish this specific video that we generate with this automation to our social media. 14:00.320 --> 14:04.040 So for an example, TikTok YouTube shorts here, Facebook, LinkedIn. 14:04.480 --> 14:06.080 For that we are using potato. 14:06.240 --> 14:09.000 So yeah you can you can just use this application. 14:09.040 --> 14:15.440 It's the most it's the best application so far for publishing the content from Nadine for me on social 14:15.440 --> 14:15.960 media. 14:16.200 --> 14:21.040 So yeah, I explained the entire process here in this video probably or even in other video. 14:21.320 --> 14:23.520 And for now I can say thank you for watching. 14:23.520 --> 14:27.160 I hope you enjoyed it and I will see you in the next material.