A guide to the various online AI image generators

A guide to the various online AI image generators

Giovanni Blandino Published on 10/28/2024

Some people are comparing the recent dawn of artificial intelligence with the invention of printing or the 20th-century digital revolution. As with all major innovations, nobody can be sure exactly where AI will take us. What is clear, however, is that generative AI is breaking new ground, making lots of previously rather cumbersome processes very straightforward, and providing an array of handy tools for various sectors. And the world of graphic design is no exception.

Creating AI images is one of the biggest opportunities generative AI provides. Today we’d like to take a better look at what artificial intelligence can now do with images and the best tools for creating AI images online, focusing in particular on those available free of charge.

But first of all, a short introduction is required, for the benefit of those approaching this world for the first time.

What can AI do with images?

Let’s start with the basics: what is generative AI? Generative AI is a type of artificial intelligence that creates new digital ‘things’, such as images or text. To be precise, generative AI creates new data by learning from existing data. It follows basic instructions provided by the user in natural language (technically known as prompts or text commands). With the best tools, all you need to do to get an image is write what you want, for example:

“A lighthouse on a rocky cliff. A stormy sea below. Photographic style”. 

And this is what you get:

Images generated by the free AI tool Adobe Firefly.

These tools’ basic function is therefore creating images from scratch via a simple written description.

But AI image generators are extremely powerful engines, and the tools they incorporate are starting to be integrated into all the leading graphic design software on the market. As a result, AI can also help to speed up retouching and many of the other jobs done by those working in the computer graphics sector.

Here is a summary of just some of the things AI image generators can do:

  • Create an image from scratch via a text command (text-to-image)
  • Create an image from a sketch (image-to-image)
  • Edit an existing image by deleting a particular detail or replacing it with something else. For example, replacing a gazelle with a lion in an image of the savannah.
  • Expand an existing image by creating new details around the original image that are consistent with it
  • Create patterns and textures within the image
  • Remove an image’s background
  • Enlarge an image without affecting its resolution
  • Improve the focus of a photograph

And much more besides!

As you can see, professional graphic designers and enthusiastic amateurs alike now have infinite scope for experimenting with artificial intelligence and images!

AI-powered graphic design software

The makers of the most popular graphic design software, including free programs that can be used online like Canva and Adobe Express, immediately twigged the power of artificial intelligence, meaning whichever software you use, you are sure to see new AI-powered functions popping up.

Today, we like to take a look at some of the best AI image generators currently available online. We’ll show an example of what each one can do and clarify whether they are free and their terms of use.

  • Magic Studio in Canva
  • Adobe Firefly
  • Image Creator in Microsoft Designer
  • DALL-E 3
  • Midjourney
  • Stable Diffusion
  • ImageFX from Google

Let’s get started!

Canva’s AI image generator

Canva is an easy-to-use online graphics program that allows users to create graphic design products and layouts for both printing and digital purposes in just a few clicks: social network posts, flyers, posters, T-shirts, mugs, calendars and much more.

Canva has now introduced an AI image generator to its software, to help users create and edit graphic design products. All of Canva’s AI-powered functions are found under the Magic Studio banner.

Canva can generate new images from a text description (text-to-image) in your chosen style: photographic, 3D, painted and many more. You can also use artificial intelligence to retouch images (for example adding new elements), remove details (using the magic eraser tool) and correct images (such as improving the focus of a blurred image).

Let’s look at a practical example of AI retouching. let’s say we want to replace the hat in the background of this photo with a Christmas tree. Using the Magic Edit tool, we select the object we wish to replace and write that we want a Christmas tree there instead. Here is the result:

Canva’s Magic Edit tool. Image:canva.com

One of the main benefits of Canva is that you can choose whether to use its proprietary AI or a third-party tool like DALL-E, Google’s Imagen and many others, meaning that all AI-powered features can be used with Canva.

Is Canva’s AI image tool free?

Canva is partially free. Its free plan includes some restrictions on the use of artificial intelligence. For example, you can only use the text-to-image tool up to 50 times in total, while subscription buyers can use it up to 500 times a month.

Adobe Firefly: Adobe’s web app with AI

Firefly is Adobe’s AI image tool. It is integrated into various Adobe products, including software aimed at professionals like Adobe Photoshop and Adobe Illustrator.

But it is also included with Adobe Express, the firm’s free online graphic design software that allows users to create graphics and layouts in the blink of an eye.

So how do Adobe’s AI image tools work? Firefly has its very own generative AI model called Image 3. You can create new images from scratch simply by describing them in text format, and there are also various tools for editing images:

  • Generative Expand: have you created too small an image? This tool allows you to expand an image’s borders, creating a new image that blends seamlessly with the existing one.
  • Generative Fill: you can use a brush to fill the image with new details and delete others, while maintaining a coherent overall effect.
  • Structure Reference: it can be a challenge to write text prompts for the image you want to create. This tool allows you to feed the AI with a reference image, so it can generate its own images based on yours.

Let’s take a look at the Generative Expand tool in action. If you want to change the shape of an image or zoom out to provide more context or add elements to your story, Generative Expand can help. Below is an example done in Photoshop, using the Firefly AI image software.

When you widen the background, the AI immediately generates new content that blends naturally with the existing image. You can then improve the new content via text prompts.

Is the Adobe Firefly AI image tool free?

There is currently a free version of Adobe Firefly that allows you to experiment without splashing out on the paid version, but you are limited to 25 generative credits (essentially 25 uses of AI, click here for details).

Image Creator in Microsoft Designer

Image Creator is Microsoft’s free AI image tool, and uses OpenAI’s DALL-E 3 engine.

This simple and fast software focuses predominantly on text-to-image: as soon as you land on the home screen, the AI prompts you to write a detailed description of the image you want to create.

For best results, the description must be very detailed and also include information on the style you wish to use, particularly if you’re creating images for a brand and so need to be consistent.

Image Creator does not currently offer more sophisticated tools, but it may prove to be the best choice of AI image tool for you, as it is completely free and very straightforward to use. It is invaluable, for example, if you need to quickly create an image for social media, produce some sketches to show to customers or draw on image generation to come up with new ideas for your latest project. Image Creator is also integrated into the Microsoft Edge web browser.

Is Microsoft’s Image Creator AI tool free?

Image Creator is totally free and you can generate an unlimited number of images. However, to access it you need to register and create a Microsoft account.

DALL-E 3, ChatGPT’s AI image tool

DALL-E is the AI image tool from OpenAI, the company behind ChatGPT. Now on its third version, the most powerful yet, DALL-E is integrated into ChatGPT, but you need to be on the payment plan to access it.

DALL-E’s main focus is text-to-image. Using the ChatGPT interface, you can ask artificial intelligence to create complex images, and improve them or change their format one text command at a time.

The key is to come up with very detailed and precise text commands that also include the style to use and potentially even the number of objects to depict and where to put them. Here is an example of a meticulous text command for generating an image, where the user has specified the breed of dog to depict, the type of clothing, what the character is doing and the various colours required.

A detailed text command in DALL-E 3. Image: zapier.com

DALL-E has recently added some tricks to help users edit their newly created images. If you select the image, you can highlight the area you wish to change, allowing you to insert details, change the background or alter an object in the image more precisely.

Is the DALL-E 3 AI image tool free?

Not at the moment: to access it you need to be on the $20/month ChatGPT Plus payment plan. However, you can test it out free of charge using the Bing search engine and Microsoft’s Image Creator, albeit without the more advanced image correction tools.

Midjourney

Midjourney is a powerful generative AI image tool, and one of the first platforms to make creating realistic images universally accessible.

Currently, the only way to access Midjourney is through the Discord platform, which does not have a free plan.

The Midjourney interface on the Discord platform. Image: docs.midjourney.com

The language (text prompts) used by Midjourney is slightly more complex than the other platforms, but it allows you to create truly professional and stunning images, so it is worth investing some time in getting to grips with it.

The example shown above has a very simple text prompt: “vibrant California poppies”. Each command generates four images which you can either discard or choose to further develop. You can, for example, zoom out, leaving the AI to imagine the surroundings of the starting image. You can change the shape, for example from square to rectangular. And you can choose whether to alter the entire image, even if only slightly, or whether to only change a selected area. You can also merge two different images using the blend tool.

Here’s an example of the blend tool in use. The AI was asked to combine the two images on the left.

Combining two images in Midjourney. Image: midlibrary.io

Is the Midjourney AI image tool free?

Midjourney has alternated between periods of being free and requiring a subscription. It currently requires payment: users can choose from four plans, the cheapest of which costs $10/month (with a discount available for an annual subscription).

Stable Diffusion: AI for free and open-source images (for research purposes)

If you are looking for a free alternative to Midjourney and DALL-E, Stable Diffusion may be right up your street.

Stable Diffusion is free (for the non-commercial version) and completely open source. This means that the models it is based on are accessible to anyone to study or improve. Another difference with Stable Diffusion is that it can be installed on your PC and used locally, or through web interfaces like AUTOMATIC1111.

Stable Diffusion is an AI text-to-image generator. This means that its main purpose is generating images from a text command provided by the user. But – thanks to a community of passionate volunteers – this already very powerful tool is constantly evolving. Stable Diffusion also offers a tool called ‘img2img’, or ‘sketch’, for image-to-image generation: you provide a sketch of the image you want and ask the AI to recreate it in a photorealistic format. Here’s an example:

The sketch function: from a sketch in Paint to a more photorealistic image. Image: medium.com

On the other hand, it has to be said that Stable Diffusion is quite complex to use: best suited to those who want to go into some depth and have plenty of time to learn how to use it and experiment with its features. If you are looking for better usability and an easier-to-use online platform, while retaining Stable Diffusion’s power, you would be better off going to one of the numerous sites that use it as their ‘engine’, like Leonardo.ai.

Is the Stable Diffusion AI image tool free?

Currently the Stable Diffusion AI image tool is essentially free of charge. The Leonardo.ai platform, which uses the same AI but with a more intuitive interface, also has a free plan with a limited number of daily uses.

ImageFX, Google’s experimental artificial intelligence (by invitation only)

It goes without saying that giants like Google are also experimenting with using artificial intelligence with images. The owner of the most popular search engine in the world has ImageFX, an online platform based on the firm’s proprietary Imagen AI system.

At the time of writing, ImageFX is not available in the UK or EU, but you can preregister for the service by leaving your email address. It’s probably worth your while, as it looks like a free and easy-to-use tool in Google’s trademark style.

Once you have entered a text command, Google creates drop-down menus for each word to help the user change the details in the image. In the example shown below, when you click on ‘pastel colours‘ Google suggests interesting variations like neon colours, dark colours and light colours.

The Google ImageFX AI image creation platform. Image: zdnet.com

Is the ImageFX AI image tool free?

Google’s ImageFX is totally free. Unfortunately, at the time of writing it is only available in a limited number of countries, and the UK is not one of them. However, you can leave your email on the website to be notified as soon as the platform is accessible.

How about you? What is your favourite AI image tool? Do you have any other platforms to tell us about?