Midjourney Tutorial: How to Generate Images Using AI and Text Descriptions

Midjourney is an AI program that utilizes cutting-edge technology, potentially based on Stable Diffusion, to generate images from text descriptions, much like DALL-E and Stable Diffusion by OpenAI. Led by David Holz, co-founder of Leap Motion, Midjourney has entered an open beta phase that began on July 12, 2022, and has already turned a profit, according to Holz. Through specific commands, users can interact with Midjourney’s Discord bot to create images, making the platform an innovative research laboratory.

To access Midjourney, you must first create a Discord account. Discord is a chat application. It has gained immense popularity among various hobby and art communities. Midjourney leverages Discord’s capabilities, although the process may seem slightly complex. If you do not already have a Discord account, you can visit their website and register for free to get started with Midjourney.

After creating your Discord account, you can proceed to sign up for Midjourney by visiting their website and clicking on the “Join the Beta” button. This will prompt an invitation to join the Midjourney Discord channel, which you can accept to gain access to the platform.

To enroll in a Midjourney plan, navigate to one of the designated newcomer rooms such as #newbies-28 or #newbies-24.

Enter the command “/subscribe” in the message field and press enter or return. This is known as a slash command, and it’s the method of interacting with Discord bots like Midjourney. By using this command, you are informing Midjourney of your intention to subscribe.

After that, a link will appear that will direct you to a page where you can sign up for a paid Midjourney plan.

Midjourney’s pricing plans begin at $8/month for the Basic Plan, which allows approximately 200 image generations per month. Standard Plan at this $24/month and The Pro Plan, on the other hand, costs $48/month and offers roughly 1,800 AI-generated images. After choosing the plan that best suits your needs, return to Discord to begin using Midjourney.

Create your first image using Midjourney.

From this point onwards, you just need to follow the prompts given to you by Midjourney. Make sure to be very specific in your instructions and avoid overloading the AI with too many details. The Midjourney bots are quite capable of handling the job satisfactorily.

After the initial image grid has finished generating, two rows of buttons appear:

U1 U2 U3 U4

U buttons upscale an image generating a larger version of the selected image and adding more details.

V1 V2 V3 V4

V buttons create slight variations of the selected grid image. Creating a variation generates a new image grid similar to the chosen image’s overall style and composition.

🔄

The 🔄 (re-roll) reruns a Job. In this case, it would rerun the original prompt producing a new grid of images.

After using upscaling an image, a new set of options will appear:

🪄 Make Variations  Web ↗️  ❤️ Favorite

Make Variations: creates a variation of the upscaled image and generates a new grid of four options.

Web: Open the image in your gallery on Midjourney.com

Favorite: tag your best images to easily find them on the Midjourney website.

Midjourney utilizes high-performance Graphics Processing Units (GPUs) to interpret and process each instruction. When you subscribe to Midjourney, you are essentially buying GPU time. Every time you generate an image, you utilize some of your subscription’s GPU time. You can use the /info command to check your Fast Time Remaining and see how much of your trial time is left.

To save your generated image, first click on it to open it to its full size. Then, right-click on the image and select “Save image.” On a mobile device, long-tap on the image and tap on the download icon located in the top right corner. It’s worth noting that all images are available to view on midjourney.com/app immediately after they are generated.