Using Midjourney for Creating Custom Web and Mobile Design Elements

Artificial Intelligence (AI) has emerged as a force to be reckoned with and has revolutionized industries and reshaped the way we work and live

AI-powered technologies like ChatGPT and Midjourney are at the center of this technological revolution and are utilized by many. These advancements in natural language processing and machine learning impact our professional lives and influence us both at work and outside of it.

It seems to be crucial in today’s world to keep up with new technologies. Learning how we can incorporate them into our daily work can help us ensure improved efficiency and support our creative processes both in our professional and our day-to-day lives. Because of the sudden growth of AI tools, such as Chat GPT and Midjourney, many professionals have started raising concerns about whether those innovations will give us new possibilities and advantages or disrupt our work lives and bring more harm than help. As of now there are two sides, the one that believes the latter to be true and the one that states that utilizing those solutions will get you ahead of your competition. If you’d like to gain that advantage, then this article is for you as it goes into detail on how to make use of one of the newest AI-powered technologies, Midjourney.

What is Midjourney?

Midjourney is an AI that generates images from “prompts”, commands made in natural language that explain what we’d like to see in the picture. It’s one of the first AI models of that sort and it was first initiated on the 12th of July 2022 by David Holz.

Currently, Midjourney, DALL-E, and Stable Diffusion are the most popular AI models generating images.

Midjourney Use Cases

What can you use such AI for? Mainly, to create amazing AI art for your pleasure. If you’re someone who likes to look at art of different kinds and can spend hours going through images found on the Internet, then this tool is perfect for you. But beware, generating images can be highly addictive.

Secondly, this tool can also be used to entertain others, for example, by using it to create memes. Below you’ll find an example of the usage of the command ‘blend’. You can take two completely different pictures and “merge” them to create something in between.

You can also use this tool for creating game assets.

Midjourney is pretty good at creating paper materials for RPGs, such as world maps, battle boards with grids, location illustrations, NPCs, monsters, etc.

It can also help with graphic projects. It won’t design perfect, ready-to-use graphic projects, since it has some limitations and is not good with things like subtitles and headlines (it generates something that looks like letters, but won’t create real words). However, that won’t be a problem for graphic designers. They can take a picture like this and create the writings or correct something by themselves.

It may also help with creating a logo. Even though the same problem as with graphic projects occurs — it cannot create proper headlines or brand names — the designs generated by Midjourney can serve as inspiration. It can also create postcards and T-shirt prints, and help with product design and interior design.

Midjourney can also be used for UI design, but only for inspiration purposes as it doesn’t create readable text and doesn’t understand the context of the project. So, can Midjourney replace UI designers? Definitely not. It is always recommended to seek the expertise of an experienced UI/UX design company when developing an app. However, Midjourney is a perfect tool for visual exploration and one can take inspiration from the findings for the page layout, type of illustrations, color palette, and other things of this sort. It can be a pretty good point of reference when starting our work on user interface designs.

Visual Inspiration

One of the aspects in which Midjourney can come in handy is its inspirational purpose. Anyone can go on Midjourney’s page and look through all the creations posted by other people. It’s not only a lot of fun to go through the pictures created by others, but it also can serve as a creativity booster. To show you just how useful it can be, let’s dive into a commercial project that I’ve done for Profil Software with the help of Midjourney. It was a project for a telemedicine app called “Appatient”, and I was tasked with creating a landing page for it’s website.

Profil Software’s project — Using Midjourney for App UI Design

Finding inspiration for the project was the first step in my work. I started by searching for any ideas about how the illustration on such a landing page could look. I typed “Isometric UI Floating Elements” and was shown some interesting results that seemed to be quite aesthetically pleasing from afar, but if one looked at them closely, it was easy to see that they didn’t make much sense. Often, there was nothing I could use on closer inspection, but the general aesthetic was quite good.

What is pretty cool about Midjourney is that you can see the prompts that others used and what parameters were applied. You can also check the original grid of the image and if it was generated using previous generations or image prompts, as well as look through the history of its parent images. There is an option that enables you to copy the whole command that was used by the other person to obtain similar results, but modify the prompt to meet your needs.

So, how did I use an AI image like this for my purposes? I was inspired by a very simple design proposed to me by Midjourney, and I decided to check what it would look like on my landing page. That’s why I downloaded the picture and cut out the background.

The next step was to test it as a placeholder on my landing page. As you can see the result is quite satisfying. Having in mind my previous independent experiments with the images for the landing page, I came to the conclusion that this one was actually a good start.

I set on the path of trying to create AI art that would be even more useful for my purposes.

I tried generating something that would meet my needs. To do that I checked what Midjourney UI/UX design prompts other people generated and I added a word “telemedicine” to one of my prompts because, like I said before, that’s what my project was about. Some of the medical results I got were a little disgusting so I couldn’t use them. Below you’ll see some of the prompts I tried:

Prompt: An isometric UI interface, bold, simple, clean, cel-shaded, plain background, white background, white, blue, light green

Prompt: smartphone, isometric, UX-UI design, 3D illustration on white background, telemedicine

As you can see, I went through a bit of an experimental phase where I tried different prompts and looked through the results in the search for something that I could use.

As I went through all of my generated images, I concluded that I liked the first one the most. That’s why I decided to use my initial inspiration for an image prompt. I entered this prompt:

Prompt: <http://s.mj.run/Y_yNhwnqnTY> isometric, telemedicine app 3D UI elements, simple

I used a link to the image for the Midjourney to take inspiration from.

And I was rewarded with a grid. It looked closer to what I wanted to achieve. The third picture in the grid especially looked like something I might want to use, so I created a few variations of it with a variation button below the grid:

Here, I focused on what elements would be easy to reproduce in Illustrator in vectors, so that I could use it in the final image and easily animate it as well. I chose the variation that was the most optimal and upscaled it for graphic processing.

The elements here are a little clearer, which gave me the possibility of recreating the most interesting elements in vectors and adjusting them to my project needs by using real icons related to the subject and inserting a picture of a doctor.

The rest was solely my work, no Midjourney in this part of the process. I ran the vector illustration through Adobe XD to organize the layers and prepare them for animation. Then I put them in After Effects and animated some of the elements.

The animation of elements in After Effects was done in a way that enabled them to be looped and rendered in json format.

Below you’ll find the final result: an image of an animated illustration for the landing page with some of the elements rotating around a glinting mobile screen.

How does it work?

I already showed you the results of my Midjourney exploration, but now let me give you some tips on how to start using it, how to create good prompts, and what is important if you want to create something useful or entertaining with Midjourney.

To start using Midjourney you’ll need an account on Discord. To start a free trial you need to enter one of the newbies-# channels. There, you’ll see active users and their work. The disadvantage of this option is that you’ll have to look through the works of others to find yours. You can also buy a subscription, which lets you generate works in the private chat with Midjourney Bot. Thanks to this you’re the only one who can see your work.

To interact with Midjourney Bot you’ll have to enter commands that start with “/”. Commands are used to create images, change default settings of the bot, monitor user information, and other helpful purposes. The most important command that is needed to create images is: /imagine. In the “prompt” field, write whatever you want to see in the picture. If you want to know more about prompts and how to use them, Midjourney offers the whole guide on their site. It will help you get started: https://docs.midjourney.com/docs/quick-start and create better prompts: https://docs.midjourney.com/docs/prompts-2.

Tips on how to create good prompts

Prompts can be very simple. Single words (or even emoticons!) will create an image. Very short prompts will rely heavily on the default Midjourney style, so more descriptive prompts will be better for a unique look. However, very long prompts are not always better. It’s best to focus on the main concepts you want to create.

One thing that you need to remember when generating images with Midjoruney is that the Midjourney bot doesn’t understand grammar, sentence structure, or words like humans do, which means that your choice of words matters greatly, especially if you want very specific results. More specific synonyms work better in many cases, so what Midjourney recommends is to, for example, use “gigantic”, “enormous”, or “immense”, instead of “big”.

You should use as few words as possible too. Fewer words mean each word has more impact. Don’t forget to put in commas, parentheses, and hyphens to organize your thoughts. However, you need to realize that the Midjourney bot will not interpret them reliably. The Midjourney Bot doesn’t account for big letters either.

It’s also better to describe what you want than what you don’t want. If you ask to have a party “without a cake”, your image will probably still have cake in it. If you want to make sure that the final image won’t have that cake, you should try the prompt with the --no parameter.

Whatever hasn’t been said may surprise you in the end. Be as specific or vague as you like, but anything you leave out will be chosen randomly. Being vague is a great way to add variety, but you may not get the specific details you want.

Where do we stand and where do we go from here?

There are many memes revolving around AI capabilities. A year or so ago people were laughing at how stupid AI generated images could look like. Below you’ll find an example:

However, not long after it was launched, Midjourney was able to generate pictures of hands and people that looked like this, which only shows how wrong we were about it:

If we reflect on the future of Midjourney and its impact on society, we will quickly come to realize that this particular AI model has broadened the horizons of creativity and convenience. Its ability to generate images based on natural language prompts will in all probability revolutionize various industries, such as art, graphic design, gaming, and even product design. Actually, one may say that it’s already revolutionizing them. However, even though the capabilities of this AI are pretty impressive, there is still a long road ahead of it before it’s able to work at a similar level to a human being.

Midjourney’s influence on society goes beyond the professional environment, since it provides users with the means to influence society and manipulate the media. It’s been already used by many to create fake news such as Trump’s arrest and other wildly spread news.

It can also be used to explore artistic inclinations, create memes, and generate content for personal enjoyment.

Midjourney is completely free, very easy to use and has a broad range of commands and parameters which makes it accessible to both beginners and professionals. Additionally, features like blending images and generating image variations bring even more innovation and versatility to this AI’s capabilities, and the images it generates are highly detailed.

With Midjourney’s continuous development and growing popularity, it is clear that AI-generated image creation has become a helpful tool for various industries and individuals. As technology progresses, we can anticipate further advancements and improvements in the capabilities of AI models like Midjourney, paving the way for a future where human creativity is augmented by intelligent machines, leading to new possibilities and exciting creations.