AI Video Creation Workflow — Step-by-Step Guide & Transcript

[00:02] starring vegetables is getting millions of views on YouTube. And in this tutorial I'll show you the workflow that many creators use to efficiently produce similar content.

[00:18] Furthermore, with the latest Google Flow tools, it is possible to maintain a high level of visual consistency between characters, settings, and voices, which obviously makes it easier to create more professional and engaging stories for the

[00:34] audience. Although results vary depending on content quality, consistency, and channel strategy, this method can help you create publish-ready videos and scale your content production. And if you want my

[00:48] cousin Franco to hold your hand through live video calls, pre-recorded classes so you can go at your own pace, personalized audits of your content where I tell you what to do, what to add, what to remove or what to change

[01:02] so you get results and can live off a completely automated business. I invite you to my new mastermind/digitalempirescommunity. description of this video so you can enter right now and I'll see you

[01:18] inside. For you, I have already created this master prompt for video generation. I'll leave all the commands in the description for you to use. Simply look for where it says to copy and paste the master commands or prompts that I

[01:34] used in this video. You're posting them so much in the Cloud AI chat because we're going to use Cloud as artificial intelligence. And as you can see, I'm artificial intelligence. And as you can see, I'm only using the free plan.

[01:48] Once you press send, it will ask you to enter the language in which you want to create videos in Spanish, since this type of content is mainly consumed by Latin American audiences. Next, it

[02:02] will generate 10 video ideas. Simply enter the number of your preferred idea to continue. It will then generate multiple scenes for a generate multiple scenes for a long video of between 3 and 5 minutes using

[02:16] only dialogue. You'll see that each scene includes the scene number, duration, details of what happens, and the characters' dialogue. If you 're satisfied, simply answer yes to lock the script and generate

[02:30] character prompts. Then, it will generate detailed text-to- image prompts to create all the character images. For each character you will get name, file name, vegetable type, role in the story,

[02:46] personality, clothing, voice note and finally a detailed text- to-image prompt. Simply copy the text prompt to the image of the first character and

[02:58] visit Google Flow. You can find the link either by searching on Google or by simply typing "Google Flow" to bring up the website. Open a new project and in the left sidebar you will see several options. Go to the

[03:15] character option. Paste the detailed prompt there, choose your preferred model, and in a few seconds your first character will be created. Then go back to your first character will be created. Then go back to Cloud AI and copy the filename

[03:30] of the first character. Replace Untitle character or untitled character if you are doing it in Spanish with the copied name. Now comes the most exciting part, voice consistency. You can assign a voice to each character

[03:45] to maintain 100% consistency. Listen to the available voices and choose the one that best suits your character. Repeat the exact same process to create all the characters and assign them a suitable voice. As you can see, I

[04:01] have already generated the five characters. Then answer yes again and this video generation engine will create all the environment prompts and scenarios that we will use to maintain environmental consistency.

[04:16] Each prompt will tell you the environment name, the file name, the scenes where the time of day appears, the key elements, and a detailed text-to-image prompt for that environment. Next, copy the prompt from the first

[04:30] environment and look at the scene option in the left sidebar. Make sure the the 16x9 aspect ratio image option is selected. Very

[04:42] important, paste the prompt and in a few seconds the environment image will be ready. Then go back to cloud AI, copy the file name, and rename the image with that name. Follow

[04:55] the exact same process to generate and rename all environment and scene images. As you can see, I have already generated and renamed all the environment images. Now go back to Cloud AI and answer yes once again. It will generate

[05:12] text-to-image prompts for all scenes. Each prompt will include information such as the characters present, the environment where the scene takes place, and a detailed text-to-image prompt. For

[05:25] example, in scene one there are two characters, Spotty Hero and Carry Love, and the scene takes place in Village Lane or whatever you want to call it in Spanish. And this happens in the morning. Now go to Flow and enter the scenes section.

[05:42] In the prompt box type @ and select the image of the first character. Make sure you choose the version labeled as image, not character. Add it to the prompt. Repeat the process with the second character.

[05:56] Then return to Cloud AI to verify which environment is being used in the scene. You will type 'a' again and select the corresponding environment image

[06:08] . Finally, paste the prompt, the text prompt, I mean text to image of said scene and in a few seconds you will have the image of the scene ready. Then you rename the image with the scene number. Follow

[06:22] exactly the same procedure for all other scenes. It's more of the same. Select the appropriate characters, the suitable environment, paste the prompt, and rename the image once it has been generated. As you can see, I have already

[06:35] generated and renamed the images for all the scenes. Then he returns to Cloud AI and answers again, "Yes, generate image-to-video prompts." If

[06:47] you wish to change the video language at this point , you can also do so. If you want to generate the video in a language other than the one selected above, enter a prompt to convert the dialogues to the desired language. However,

[07:00] for this example we will continue with the language already selected, which is Spanish. Once you press send, it will generate picture-to-video prompts for all scenes. Each prompt will include the image of the scene to be used,

[07:15] the duration of the video, and a detailed prompt that includes the dialogue. Copy the prompt corresponding to the first scene. Go to Google Flow and select the video option. Also make sure that the ingredients option is enabled

[07:28] . To save credits you can select VI 3.1 Lite. Then you're going to type @ and select the image from scene one.

[07:40] Then type @ again and select the character who speaks in that scene. This time, make sure to choose the item labeled as character, not image. This will help maintain the character's voice consistency across

[07:54] all scenes. Next, paste the video promo and in a few seconds your clip will be ready. You can preview both the video and the dialogues, and follow exactly the same process for all other scenes. Copy the prompt,

[08:10] , paste the prompt, and download the video once it has been generated. As you can see, I have already generated all the video scenes and downloaded them. Now open your favorite video editor, import all the clips and place them on

[08:26] the timeline in the correct order. Make small edits by removing any unwanted parts, and once finished, we must remove the watermark. To do this, add a surface blur effect and extend it

[08:41] to the end of the timeline. Preview the result and as you can see, the watermark is no longer visible. Then go to the transitions section and add a fade transition between the clips by setting a duration of

[08:54] the clips by setting a duration of 0.5 seconds or 15 frames and repeat this process for all the necessary clips. When everything is ready, perform a final check before exporting. [laughs]

[09:08] Carry, today we'll sell all our potatoes and carrots. Together we always succeed, Spudy. And that's a fact. We are the best team in town. We are.

[09:23] are. We've worked together since we were little. I will never leave you alone, Carry. And once you're satisfied, you export the project. Now it's time to make it go viral by publishing it with

[09:36] SEO-optimized metadata. Go back to Cloud AI and answer yes, but make sure to request the title and description in a combination of your selected language and English. In our case, the language is Spanish. It will generate three

[09:52] title options. An SEO-optimized description for virality, as well as the best tags to maximize reach. It will also provide a suggestion for a pinned comment and a detailed text-to-image prompt for generating

[10:05] a thumbnail. Use that prompt in Google Flow and you'll get a thumbnail with a high CTR (click- through rate) so that people, your audience, can click on it as soon as they see that thumbnail. If you already have a channel

[10:18] ready to publish these videos, you can proceed directly and publish the content. But if you want to create a channel from scratch optimized for this type of video, pay attention. I have also created a master prompt, specially

[10:31] designed to configure a new channel optimized for this type of content. As I said, you simply copy the prompt, open a new chat in Cloud AI and paste it in full. It will ask you for the language of the video you are about to publish.

[10:45] Enter it, then it will suggest some of the best channel names along with usernames. Copy the username from youtube.com followed by that username and press enter. If a page appears indicating that it is unavailable,

[10:59] it means that the username is free and you can use it. And once you 've chosen the channel name, reply with the corresponding number. It will generate a description optimized for SEO and especially virality, as well as

[11:11] tags for the channel. Answer yes once more and it will provide you with a detailed text-to-image prompt for your profile picture. Use that prompt in Google Flow to generate the image, then answer yes again and it will generate

[11:27] a detailed prompt for the channel banner, i.e., the cover photo at the top . Use that prompt in Google Flow to also create the banner. And that's exactly how you can easily create and scale a channel with

[11:42] this type of content. By the way, if you want me to show you how I'm creating, growing, and scaling automated YouTube channels from $0 to $3,700 per month in just 30 days, check out this video that's on your screen. And if

[11:57] you want to see how my client earned $4,400 in just 30 days by applying exactly what I teach, watch this other video that appears on your screen. other video that appears on your screen. Check it out.