Midjourney Tutorial

Step 1: Sign Up and Join the Midjourney Discord

First, go to midjourney.com and click 'Join the Beta'. This will redirect you to Discord. If you don't have a Discord account, you'll need to create one—it's a quick process. Once logged into Discord, you'll be invited to join the official Midjourney server. Accept the invite. The server can be overwhelming, with dozens of public channels where users are generating images non-stop. Don't try to work here. Instead, find the Midjourney Bot in the member list on the right (its icon is the Midjourney logo). Click on it and select 'Message'. This opens a private Direct Message (DM) window with the bot. In my experience, this is where you should do all your work for privacy and focus. However, note that you must generate a certain number of images (usually 25) in public channels before the bot will respond to you in DMs on the Basic plan. So, for your very first images, you'll need to find a 'newbies' channel in the main server.

Step 2: Understand the Discord Interface for Midjourney

Midjourney has no traditional website dashboard; your interface is Discord. In your chosen channel or DM, you'll type all commands. The core command is /imagine. Simply type a forward slash (/) and then start typing 'imagine'—Discord will auto-suggest it. Click on it, and a prompt box will appear. This is where you describe your image. Below the chat window, you'll see the bot's responses. When you submit a prompt, the bot generates a grid of four low-resolution image options. Underneath this grid, you'll see buttons labeled U1, U2, U3, U4, and V1, V2, V3, V4. The 'U' buttons are for 'Upscaling'—selecting one image to generate a high-res version. The 'V' buttons are for 'Variation'—creating a new grid of four images based on the chosen thumbnail's style and composition. This button-based interaction is the heart of the Midjourney workflow.

Step 3: Craft Your First Prompt and Generate Images

Let's create your first image. In the prompt box that appears after typing /imagine, describe what you want to see. Start simple. I tested countless prompts, and beginners often overcomplicate them. Try something like: 'a cute corgi puppy sitting in a field of sunflowers, cartoon style'. Then press Enter. The bot will acknowledge your job and, after about a minute, post a grid of four images. What surprised me was how much detail it adds from such a simple prompt. Now, look at the four options. Choose the one you like best. If you like the general idea but want to see similar alternatives, click the corresponding 'V' button (e.g., V2). If you love one and want it in high resolution, click the corresponding 'U' button (e.g., U3). The bot will then generate that single, detailed image.

Step 4: Use Parameters to Take Control

This is where Midjourney gets powerful. Parameters are special commands added to the end of your prompt to control the output. They always start with two hyphens. The most important one for beginners is --ar for aspect ratio. The default is 1:1 (square). To create a widescreen landscape, add --ar 16:9. For a portrait, use --ar 2:3. Another crucial one is --v followed by a number (like --v 6.2) to specify which Midjourney model version to use. I always use the latest version for the best quality, but sometimes older versions have a specific aesthetic. You can also use --s for stylization (higher values make the AI more 'creative') and --chaos (higher values, from 0-100, create more diverse and unexpected grid options). Add them to the end of your prompt: '/imagine a majestic fantasy castle --ar 16:9 --v 6.2 --chaos 30'.

Step 5: Upscale, Save, and Find Your Masterpieces

Once you upscale an image (by clicking a 'U' button), you get the final high-resolution version. To save it, simply click on the image to open it in your browser, then right-click and 'Save image as'. On mobile, press and hold. But here's the critical part I learned the hard way: Discord is a chat stream. Your amazing images will quickly scroll away and be hard to find. You must use Midjourney's dedicated website gallery. Go to midjourney.com and log in with your Discord account. Click on 'Community Feed' to see public images or, more importantly, click on your profile icon and select 'My Images'. This is your personal, searchable archive of every image you've ever generated. You can favorite them, download them, and see the exact prompt used. Never rely on scrolling through Discord to find your old work.

Step 6: Explore Advanced Prompting and Features

Now that you know the basics, let's unlock more power. First, learn about multi-prompts. Using double colons :: gives different parts of your prompt separate weights. For example, 'sunflower field:: corgi puppy::2' tells the AI the puppy is twice as important as the field. You can also use image prompts. By uploading an image URL (you'll need to host it somewhere like Discord itself) and adding it to the start of your /imagine prompt, you can blend styles and compositions. Another game-changer is the 'Remix Mode'. Enable it with /settings and clicking the Remix button. This allows you to edit the prompt text when creating variations, giving you pinpoint control over iterations. Finally, experiment with the /describe command: upload any image, and Midjourney will suggest four possible prompts that could have created it—an incredible learning tool.