What is Midjourney and How to Use It

by | AI

Remember the not-so-long-ago days when we had to rely solely on artists to create stunning visuals?

Yeah, well, those days are over.

Thanks to Midjourney, anyone can now become an artist. All you need is a text description and your imagination! 

Midjourney is an AI-based art generator that uses interactive machine learning to create distinctive images from text prompts provided by users. The tool can create images from a spectrum of art styles, ranging from abstract to hyper-realistic. 

This grid of images was made by Midjourney in less than a minute using the prompt: “ultra-high definition, light white and light gold style, close-up straight to the face portrait, watercolor media, highly detailed, picturesque village, front lighting.”

What is Midjourney

If you want to learn how to get started with this amazing tool, read on!

This article covers everything you need to know about Midjourney, including how it works, how much it costs, how it compares to competitors and of course, its limitations. 

Let’s dive in! 

What is Midjourney?

Midjourney homepage

Midjourney is an AI platform where users can generate unique AI images based on text descriptions. It uses machine learning models to generate uncanny, hyper-realistic images in any style you could imagine. 

The company was founded by David Holz in 2021, and the service was launched to the public in an open beta in July 2022. It has since ignited a worldwide frenzy and has reportedly generated over $200 million in revenue, all without help from external investors. 

The beauty of Midjourney is its simplicity. You don’t need to download any software or buy specialized hardware to get started. All that’s needed is a Discord account and a Midjourney subscription, which starts from $10/month for the basic plan.

This low barrier to entry has made Midjourney extremely popular, with a whopping 15+ million users on its official Discord server! 

6 Steps to Use Midjourney 

Ready to use Midjourney? Great! This section of our guide will show you how.

With a few easy steps, you’ll be making amazing AI images in no time. Whether you’re new or have some experience, these directions will help.

Step 1: Log in to Discord

Sign in or create a Discord account

Midjourney can only be accessed through Discord either through a web browser, mobile, or desktop app. If you don’t have a Discord account, you can create one here. 

Step 2: Sign in to Midjourney

Authorize Midjourney to access your Discord account

Once you have a verified Discord account, head over to Midjourney.com/account and sign in. You’ll need to give Midjourney access to your Discord account. 

Step 3: Subscribe to a Midjourney Plan

Comparison of Midjourney subscription plans

Midjourney no longer offers a free preview, so you have to subscribe to a plan if you want to use it regularly.

There are four subscription tiers available. Choose one that works for your budget and needs. 

Step 4: Create a Personal Discord Server

Create your own private Discord server

Get started by creating a personal Discord server (it’s free). Click on the + button on the left-hand side of the page, then select the Create My Own option. Choose For me and my friends and then enter the name of your server and click Create

Step 5: Invite Midjourney Bot to Your Server

Add Midjourney Bot to your private Discord server

Go to the Midjourney server using this URL: http://discord.gg/midjourney. Accept the invite and press Join. Once you’re on the server, find Midjourney Bot from the User List, and then click Add to Server. Select your server from the list and then follow the directions. 

Alternatively, you can start a direct chat with Midjourney Bot. The results will be the same. 

Step 6: Use the /imagine Command

Use /imagine in discord to use Midjourney Bot

To interact with the Midjourney Bot, you have to use commands. The /imagine command is used to create images.

Type it out and then include your prompt, i.e., /imagine prompt, and then send your message. The bot will interpret your text prompt and start generating images. 

How Midjourney Works 

Community showcase of images made by Midjourney users

Midjourney interprets and processes requests using Graphics Processing Units (GPUs). When you pay for a subscription, you’re essentially buying monthly access to GPU time. 

The service offers two modes for image generation: 

  1. Fast: This is the default and highest priority tier which gives you GPU time instantly. Each subscription plan offers a limited amount of GPU hours.  
  2. Relax: This mode puts your request in a queue and can take anywhere from 0 to 10 minutes per job. Using this mode saves your monthly GPU time. This mode is only available to subscribers of Standard, Pro, and Mega plans. You can switch between modes using the /fast and /relax commands. 

The average Midjourney job takes around a minute’s worth of GPU time to create a finished image.

How to Modify Midjourney Images Using Vary and Upscale 

Once your image is created, you can create variations of a specific image using the V1, V2, V3, and V4 buttons under the grid. 

You can also choose an image to upscale using the U1, U2, U3, and U4 buttons (The numbers correspond with an image’s position on the grid).

Below is the result of the prompt: “A desert with neon-colored sand dunes under a starry sky.” You can see the buttons for variations and upscale beneath the grid.

Buttons for upscaling and creating variations of Midjourney images

We chose to upscale the first image (U1), so Midjourney upscaled it and offered options to do a strong or subtle vary, zoom in or out, or vary a region.

Options to vary and zoom in or out of your Midjourney image

Choosing Vary (Region) opens an editor where you can select specific areas you want to regenerate. Let’s say we wanted to remove the piece of rock in the corner of the image.

Portion to modify in Midjourney generated image

After choosing Vary (Region), we used the Select tool to choose the portion we want to fix and then clicked Submit.

Midourney editor for changing a region

The result was four new variations of the image without that piece of rock. We chose the best one, and this is what it looks like!

Final result of Midjourney picture after using vary (region) tool

You can use Vary (Region) to fix any type of detail. It could be a crown on an illustration of a cat or a human hand that doesn’t look quite right. It’s a very versatile tool for making quick edits without having to start over with a new prompt.

Note: All of these actions will cost GPU time. 

How Much Does Midjourney Cost?

Cost of different Midjourney plans

Due to heavy demand, Midjourney no longer offers a free demo to new users. You need an active subscription to use the service to generate images. 

There are four subscription plans: 

  1. Basic plan ($10/month): Comes with 200 fast mode minutes per month and 3 maximum concurrent jobs. This tier does not have relax mode. 
  2. Standard plan ($30/month): Comes with 15 fast mode hours per month, which is around 4.5 times more than the basic plan. It supports 3 maximum concurrent jobs and includes unlimited access to relax mode.  
  3. Pro plan ($60/month): Comes with 30 fast mode hours per month and supports up to 12 maximum concurrent fast jobs and 3 concurrent relaxed jobs. 
  4. Mega plan ($120/month): Comes with 60 fast mode hours per month and supports up to 12 maximum concurrent fast jobs and 3 concurrent relaxed jobs. 

For reference, one GPU minute is basically one image, so one GPU hour is equivalent to 60 images and roughly 200 image variations. 

If you run out of your monthly allocation of fast hours, then you can buy more at a rate of $4 per GPU hour. The extra hours won’t expire, but they require an active subscription to use. 

All images you create while subscribed are entirely yours, and you are free to use them in any way you want.  

Note: You can’t share a Midjourney subscription. The pricing plans are for individual use and are tied to your Discord account. If you invite other people to your personal Discord server, they won’t be able to use Midjourney Bot. 

Midjourney vs. DALL-E 3

Comparison of priestess tarot card made by Midjourney and Dall-e 3

Midjourney is a dominant force among AI art generators, but how does it compare against other rising stars like DALL-E 3?

DALL-E 3 is the latest version of OpenAI’s AI art generator that works right in ChatGPT. That’s right! You can use ChatGPT to brainstorm prompts and then use them to create images using DALL-E 3 right from the chat box. 

So, how do they compare?

  • Cost: DALL-E 3 is available to ChatGPT Plus and ChatGPT Enterprise users. A monthly subscription costs $20. Midjourney’s basic plan starts at $10. 
  • Image quality: DALL-E 3 generations tend to be much more accurate to the prompt, but Midjourney’s images are sometimes more visually rich. The Vary (Region) tool also gives Midjourney an edge.  
  • Ease of use: DALL-E 3 is much easier to use than Midjourney because it’s integrated right into ChatGPT. Midjourney has more steps since you can only use the bot via Midjourney’s Discord server or your own private Discord server.
  • Celebrities and public figures: Midjourney can create images from prompts containing the names of public figures or celebrities. DALL-E 3 has strict guardrails and doesn’t allow the creation of images containing public figures.

The bottom line, while Midjourney’s images may be visually rich, DALL-E 3’s generations are much more accurate to the prompt. When you combine its ease of use, DALL-E 3 is the natural winner. 

However, DALL-E 3’s censorship and content limitations are a little extreme when compared to Midjourney. DALL-E 3 also doesn’t have options to upscale or modify specific regions of images.

Overall, DALL-E 3 performs better than Midjourney in many ways, and DALL-E 3 could end up being more popular, but Midjourney remains king for now.

If you want a more comprehensive review of DALL-E 3 vs Midjourney, check out our guide! 

Limitations of Midjourney 

Content warning from Midjourney Bot in discord

Every tool has its limits, and Midjourney is no different. Like all AI image makers, it can’t do everything. This section talks about what Midjourney might struggle with. Knowing these limits can help you use the tool better. 

  • Terrible with images containing words: If you want an image with words, then get ready to get gibberish more often than not. After a lot of experimenting, we’ve concluded that the current iteration of Midjourney just isn’t great for creating images containing text. But this problem isn’t unique to Midjourney; DALL-E 3 is somewhat better at it, but it also has similar issues. 
  • Struggles with hands: Like most AI generators, Midjourney struggles to render human hands correctly. In our experiments, DALL-E 3 seemed to perform better in this regard, but the differences aren’t that big. This is an issue that’ll likely be solved in future versions. 
  • Must comply with community guidelines: This is standard stuff for AI image generators. All images you create must comply with Midjourney’s community guidelines. That means no adult content, gore, or text prompts that are abusive, aggressive, or disrespectful. 

It’s important to note that most of Midjourney’s limitations aren’t unique to the service. These are limitations of AI image generators in general, and competitors like Leonardo AI and DALL-E 3 have them too. 

Final Thoughts

Image of boy holding a bird made with Midjourney

Midjourney has opened up a world of AI-driven image magic. We’ve walked through what it is, how to make it work for you, and what it might cost. We’ve also seen where it stands next to other tools like DALL-E 3 and where it might not be perfect.

But every tool has its place, and Midjourney offers something special. Its features and ease of use make it a strong choice for many. As with any tool, knowing its strengths and limits is key.

Whether you’re diving in for the first time or weighing options, it’s clear that Midjourney has a lot to offer. It’s a tool worth exploring!

Want to learn more about how AI is transforming the tech landscape? Watch the video below! 

Frequently Asked Questions 

What does Midjourney do?

Midjourney is an AI image-generation tool that can read your text prompts and create an image based on it in less than a minute. 

What is Midjourney AI used for?

Midjourney is an AI program used for creating art and images using text prompts. It can be used to create an image in any style you desire, provided it doesn’t violate Midjourney’s community standards (no sex, gore, blood, etc.). 

Is Midjourney free to use?

Midjourney is not free to use. It requires a monthly subscription to access, and the basic plan starts at $10/month. 

What is the difference between Leonardo AI and Midjourney?

Midjourney and Leonardo have similar functionality, in that you can input a text prompt to get four images in return. However, Leonardo AI has more features and allows users to edit images, add new elements to existing images, and even train custom AI models. 

author avatar
Sam McKay, CFA
Sam is Enterprise DNA's CEO & Founder. He helps individuals and organizations develop data driven cultures and create enterprise value by delivering business intelligence training and education.

Related Posts