How Generative AI is Being Used in New Ways

Introduction

Hello everyone!
Today, I am going to share with you all what I recently found and how good it is.
Recently, I came to know about an Asset Management Platform that helps developers like us to manage assets on the web efficiently. Yes! I am talking about Cloudinary (an Image and Video API Platform). It provides dynamic APIs to manipulate assets like images and videos on the web.

But today, I am not going to share what Cloudinary is; instead, we are going to explore some new and exciting features from Cloudinary.

Quick Intro of Cloudinary

A quick introduction to Cloudinary for those who haven't heard this name before, so you can understand the blog better. If you already know what it is and have used it, you can skip this section.

Cloudinary is an Image and Video API Platform that offers dynamic APIs to host your assets on the web with good support and accessibility. Along with APIs, it also offers an Asset Management service where you can store, host, and manipulate your assets. This is how Cloudinary provides a full set of features that a developer or organization needs to ensure their service does not face high load or compromise the quality of assets for performance.

Now that you have a solid context of what Cloudinary is, let's move ahead…

Why am I writing this blog?

Now, you are wondering why I am writing this. Is this a Cloudinary doc or advertisement? No, it's not. I am writing this to share an experience with you all that excites me a lot.

Recently, I got a newsletter email from the Cloudinary team (they called it Cloudinary Developer Bytes) and I opened it. In that email, the Cloudinary team shared what new and cool things they are shipping and building. It includes various things from MCP Server to 3D Spinset to tutorials and a lot more. By going through this, I found that Cloudinary is also shipping new AI features on its platform. Honestly, at first, I thought it was some chatbots or agents to transform images or videos. But I was wrong; they really shipped some nice and cool AI features that can enhance the way we interact with assets on the web.

Here's the list of AI features for images available so far as I write this article:

You can now use Cloudinary to perform many operations on your assets directly on the web.

Exploring Cloudinary's AI Features

I tried some of these features (like Upscale, Generative Remove, Generative Replace) and I really love them:

Here are some examples:

Low Resolution Cat Image This cat image has dimensions of 340 × 400. By simply adding e_upscale/ in the URL, the image is upscaled to 1360 × 1600, almost 4 times larger. You can do this on the web by just modifying the URL parameters to upscale your image 4 times.

You can easily test it by adding those parameters between the upload and version number (Modified URL with those parameters). That's it, your image is upscaled. No website, no subscription, no tools needed.

I also tried Generative Background Replace, which lets you change the background of any image to anything you want. For example, here is an image of a girl and a dog with a solid background → Image. By adding e_gen_background_replace:prompt_times%20square in the URL parameters, it gets a Times Square background → Transformed Image.

You can also combine these effects as needed (peak customization), for example, transforming both the background and objects → Modified Image.

There are many more features available for you to explore and enjoy. And this is just for images; Cloudinary also offers various AI features for video manipulation. Check out these docs: AI In Action

How can you do this or create your own…

I faced some issues with this because it wasn't clearly available in the docs, or I couldn't find it (skill issues). However, it's not too difficult to locate. To test AI features or create your own:

Go to your Cloudinary console → Go to the image section → Go to Manage Transformations (here you will find all your saved transformations with your custom names (yes! that’s also possible)) → Then click on the + icon, and you will be redirected to the transformations console, as seen in the second image.

The Transformation Console Page:

As shown in the image, you have sample images and options to test features with images in the console. You can find the URL parameters for that feature in the URL section (you also get SDK code to use in your application). After testing, you can save your custom actions with your custom parameter name, as seen in the first image (t_abhinav).

Conclusion

I hope you find this article helpful and learn something new. If you have already used or tested these Cloudinary's features, please share your experience. Did you like it, or did you encounter any issues?

Do you know of any other tools or platforms that offer similar features? Please let me know, as I would be happy to test or use them.

Thanks for reading!

Goodbye :) See you in the next article 👋