You may already know the importance of optimising videos for YouTube search rankings especially for the content creator. This indicates that the tool that generates video titles, descriptions, keywords, and thumbnails can boost your visibility. 

However, have you ever considered creating your AI-powered SEO tool without coding experience? This blog will show you how to build a no-code SaaS tool. You’ll discover a step-by-step guide to creating a powerful YouTube SEO assistant in minutes here, from generating video metadata to thumbnail ideas.

See the video blog here

Introducing Vector Shift AI

All right to create this tool you can use a no-code SAS app builder called Vector Shift AI which is an integrated framework of no-code, low-code, and out-of-the-box generative AI solutions to build-

• AI search engines,
• Assistance chatbots, and
• Automation.

To get this too, click on this link-https://cutt.ly/7eYFjDZq once you are on the homepage of the tool, click on Get Started and then create an account. Once you are logged in your interface will look like this picture below.

Here, on the left side of the page, you will find various options. Such as, in the marketplace option, you will notice a vast number of SAS tool templates created with Vector Shift. You can also download and use them in your project or you can create your tool.

• To do that first, go to the Pipeline tab where you will find some of the pipelines created earlier.
• To create a new pipeline, click on the New button where you will find various pipeline templates. You can

1. Deploy
2. Modify and
3. Generate data using them

• Now to create a new pipeline from scratch, click on the Plus Icon and you will be driven to the pipeline Creator and editor panel.

• At the top, of the page you will find various tabs for

1. Input
2. Research and
3. Output options

• In the LLM tab, there are all the popular LLM models available. You will find options to

1. Insert a knowledge base
2. Add multimodal LLMs
3. Integration logic and others

• So, here first you have to add an input option by going to the General Tab and selecting the input node.
• Then you will get your input node. As you are entering the YouTube video URL, you need to rename it to YouTube URL.
• To convert this URL to transcribed text, you need a data loader.
• So go to the Data Loader tab to find various data loaders and select the YouTube data loader.
• Then it will analyse the YouTube video and transcribe its audio.
• Now you have to connect the input with the YouTube data loader by simply dragging and dropping the YouTube URL point to the URL to connect them.
• All right now add an LLM model to move ahead. In the LLM tab, you will find all the popular LLM models like-

1. Open AI
2. Gemini,
3. Perplexity and more

• You can use any of them for example select Claude and by this, the Claude LLM node has been added.
• Then you can see two input options there-

1. System and

2. Prompt

• Under the model tab, you can set the preferred Claude model and then you have to customize the LLM input. Before that, you need to add another LLM node. In that case, the first one is to generate the title and the second one is to generate the keywords or tags.

Now let’s see how you can do that in the blow-

• First, connect the YouTube loader output with the first LLM nodes prompt section.
• Now you have to enter a custom system prompt in the prompt box so that the LLM can understand what you want from the YouTube transcription text. You can find all the prompts along with other prompts in the Discord server and Telegram Channel.
• Now after getting the prompt, copy the Prompt, go back to Vector shift and paste The Prompt in the system prompt box to generate the video title. 

This is how the LLM is now ready to generate the title but to do that, you need an output node which you will find in the General Tab and then add one output node. Place it as needed and then rename the node according to the output. Then select the output type for example, you can go with text.

Connecting the Title Generator LLM Response

Now you have to connect the title generator LLM response to the title output’s input section. In the meantime, your title generator is ready to modify the keyword generator node.

• Go to the prompt section, copy the tag generator prompt and paste it into the second LLM system prompt box.
• Now according to the prompt, you have to enter the title and transcription input as the prompt.
• So, to create variables in the prompt box click on the icon mentioned in the picture, get your first variable created and rename it as title.

• Now, to create another variable, click there again and rename the newly created variable as transcription.
• Next, you have to connect the YouTube loader response with transcription and the title generator LLM response with the title variable.

Once done, add another output node to publish the keywords. So, add the output node, rename it and then connect to the keyword generator LLM response.

Description Generator

All right now let’s create the description generator-

• First, you have to add another Claude LLM node.
• After that enter a description generator system prompt and this time you need three variables according to the prompt.
• So, create the variables and rename them as transcription titles and keywords.
• After creating the variables, connect them with related responses.
• Now to publish the description, add another output node and rename it as description. This time set the output type as formatted text.
• After that connect with the description generator response.

All right now you are ready to generate the title tag and description.

Thumbnail Idea

So, it’s time to see how you can generate the thumbnail idea and example images.

• So, add another LLM model but this time you can use the Open AI LLM.
• After adding the Open AI LLM, enter the system prompt.
• Now according to the prompt, you have to connect this prompt with the title generator LLM response and add an output node, renaming it as a thumbnail idea.
• Then connect it with the thumbnail idea generator response.
• Now the last step of the pipeline is generating example images which requires a prompt that you can generate by image generation prompts from the thumbnail idea.
• Simply add another Open AI LLM and enter the system prompt.
• Then connect the prompt section with the thumbnail idea response.
• Now it will generate the image generation prompt.

However, you also need an image generation model that you can create by going to the Multimodal tab. There you will find various models to generate images, audio and others. You have to add the image generation node.

After that, set the image generation model and there you will find two options from where you have to choose one of them and set other settings. Once done, connect the prompt section with the prompt response node. 

Okay, now you have to add another output node and rename it. Then set the image as the output type and connect the output with the image generation node. Here you are done with your pipeline to use and click on Deploy Change to save the pipeline. 

Now if you click on the Run Icon, the pipeline will run. You can preview it by using any video URL as the input URL and hitting the Run button. You can see your pipeline start working in the background.

Deploying to Website

Now to deploy this Pipeline on your website, first you have to save the pipeline then go to the top right corner of the page where you will find an icon called Export Pipeline. Click on it and you will find three options there-

1. Automation

2. Chatbot and

3. Form

You have to select the form option and give a name to your form. After that hit the Create Form button and customize the form’s appearance on the next page. You can also add a custom logo if you want and change the name, custom message, the input and output format. Once done, hit the Deploy Change button and then hit the Export button. 

Now you will get the form published and you can copy the URL and share it with your friends. In the case of sharing it with friends, you can go with the second option by just copying the code and publishing it on your website. And this is how your newly created SAS tool has been published on your website. 

Now you can try it by entering a YouTube URL and then hitting the Generate button. You will get three titles-

1.     The video description with a minimum of 20 related keywords

2.     A thumbnail idea and

3.     In the end an image

So, Vector shift AI is more capable than you might think you can do.

Announcement

If you have a passion for a stable career, you should learn the right skills. In this case, SkillShare can be a good platform for online learning, offering thousands of classes in AI tools, marketing, design, and many more. 

It provides mastering AI tools to explore new platforms and will help you build skills that matter in today’s fast-changing world. However, you get one month of SkillShare free by using this link- https://skl.sh/ailockup07242.