How to Use MidJourney: From First Prompt to Pro-Level Control
Closeup of Smartphone screen with logo lettering of Midjourney AI text-to-image tool on computer keyboard. By Ralf
MidJourney has firmly established itself as a powerhouse in the world of AI art, celebrated for its ability to generate stunningly artistic and high-fidelity images. However, its unique Discord-based interface can be an initial hurdle for newcomers. This guide provides a comprehensive, step-by-step walkthrough of how to use MidJourney, from setting up your account and generating your first image to mastering the advanced features that provide pro-level control over your creations. By the end of this in-depth guide, you will have a clear understanding of the entire MidJourney workflow and the confidence to bring your most ambitious creative visions to life.
Getting Started: Your Pre-Flight Checklist
Before you can begin your journey into AI art creation, there are three essential prerequisites to address. MidJourney's entire ecosystem is built around the Discord platform, so a basic familiarity with it is key.
A Discord Account: MidJourney operates exclusively as a bot within the Discord platform. If you do not already have an account, you will need to create one at discord.com. The process is simple and free.
The Official MidJourney Discord Server: All interactions with the MidJourney bot happen within its official server. You must join this server to generate images. You can find the direct invitation link on the MidJourney website.
A Paid MidJourney Subscription: As of 2025, MidJourney does not offer a permanent or stable free tier for new users. To generate images, you must subscribe to one of their paid plans. This is a crucial distinction from some other platforms and is the most common point of confusion for beginners.
Joining the Server and Choosing Your Plan
Once you have your Discord account ready, the setup process is straightforward. Your choice of plan will depend on your expected usage, need for privacy, and budget.
Step 1: Join the MidJourney Server
Navigate to the official MidJourney website and click the prominent link to join their Discord server. After accepting the invitation, the MidJourney server will appear in your server list on the left-hand side of the Discord application.
Step 2: Subscribe to a Plan
Upon joining, you will be prompted to choose a subscription plan. You can also initiate this process at any time by using the /subscribe command in any channel where the bot is active. The plans are structured around GPU time, which is the server-side processing power consumed to generate your images.
Basic Plan (~$10/month): This entry-level plan provides approximately 3.3 hours of "Fast" GPU time per month. It is ideal for beginners and casual users who want to explore the platform and generate a few hundred images.
Standard Plan (~$30/month): A significant step up, this plan includes 15 hours of Fast GPU time and, crucially, unlimited "Relax Mode" generations. This is the most popular plan for hobbyists and regular users, as it allows for endless experimentation without worrying about running out of credits.
Pro Plan (~$60/month): Aimed at professionals and power users, this plan offers 30 hours of Fast GPU time and includes "Stealth Mode," which prevents your images from being publicly visible in the MidJourney gallery. This is essential for any commercial or confidential work.
Mega Plan (~$120/month): The top-tier plan provides a generous 60 hours of Fast GPU time, the highest priority access to the servers (meaning less waiting during peak times), and Stealth Mode.
"Fast Mode" generates images in about a minute but consumes your monthly GPU allowance. "Relax Mode," available on Standard plans and above, is slower (taking 1-10 minutes per generation) but does not use your Fast GPU hours, allowing for unlimited creations. For most users, the Standard Plan offers the best balance of speed and value.
Creating Your First Image Step-by-Step
With your subscription active, you are ready to create your first piece of AI art. The process is simple and iterative.
Step 1: Navigate to a Generation Channel
Within the MidJourney server, locate a channel designated for image generation. For new users, these are typically named #newbies or #general. These channels are public, so you will see a constant stream of other users' creations. As you become more comfortable, you can also generate images by sending a direct message (DM) to the MidJourney Bot. This provides a cleaner, more private workspace, though your images will still be publicly visible in the MidJourney gallery unless you are using Stealth Mode.
Step 2: Use the /imagine Command
In the message box of your chosen channel, type /imagine. This is a universal Discord command that activates the MidJourney bot and creates a prompt field. In this field, type a clear and descriptive sentence or a series of keywords describing the image you want to create. For example:
Plain Text
/imagine a majestic lion with a crown made of stars, sitting on a throne in a celestial palace, cinematic lighting, ultra detailed, 8k
Press Enter to send your prompt. MidJourney will then begin its diffusion process, starting with a field of visual noise and progressively refining it over several steps into a grid of four unique image variations that interpret your text.
Step 3: Upscale or Create Variations
Below the generated 2x2 image grid, you will find a set of interactive buttons that are key to refining your creation:
U1, U2, U3, U4: These buttons correspond to the four images in the grid (1 is top-left, 2 is top-right, 3 is bottom-left, and 4 is bottom-right). Clicking a "U" button will Upscale the selected image, creating a larger, higher-resolution, and more detailed version of that specific image.
V1, V2, V3, V4: Clicking a "V" button will create four new Variations based on the overall style and composition of the selected image. This is useful when you like the general direction of an image but want to see slightly different versions.
Reroll (🔄): This button will rerun your original prompt, giving you a completely new grid of four images.
This simple loop—prompting, selecting an image you like, and then choosing to either upscale it or create variations—is the fundamental workflow of MidJourney. It is an iterative process of discovery and refinement.
The Art of the Prompt: A Deeper Dive
While the /imagine command is simple, mastering the art of prompting is what separates a good result from a great one. The most effective prompts are not necessarily the longest; they are the most specific and visually descriptive.
Core Prompt Structure
A strong MidJourney prompt is built from several key components:
Subject: The main focus of your image. Be as specific as possible (e.g., "an ancient, weathered oak tree" is better than "a tree").
Medium/Style: The artistic style you want to emulate (e.g., "digital painting," "35mm film photograph," "watercolor illustration," "manga screen-tone").
Environment/Setting: Where the subject is located (e.g., "in a misty forest," "on a neon-lit cyberpunk street").
Lighting: This has a huge impact on mood (e.g., "soft morning light," "dramatic cinematic lighting," "glowing neon signs").
Color Palette: Specify the colors you want to dominate the image (e.g., "vibrant and saturated," "monochromatic and moody," "pastel colors").
Composition: The framing of the shot (e.g., "wide-angle landscape," "close-up portrait," "top-down isometric view").
Example of a Structured Prompt
Instead of a simple prompt like a knight, a more structured prompt would be:
Plain Text
/imagine portrait of a stoic female knight, intricate silver armor with gold filigree, standing in a sun-drenched cathedral, rays of light filtering through stained glass windows, photorealistic, detailed face, cinematic composition --ar 2:3
This prompt gives the AI specific instructions for the subject, armor details, setting, lighting, style, and composition, leading to a much more controlled and impressive result.
Advanced Techniques: Image Prompting and Seeds
MidJourney is not just for creating images from text. You can also use your own images to guide the generation process.
Using an Image as a Reference (Image Prompting)
To influence the composition, style, and color palette of your generation, you can provide MidJourney with one or more reference images.
Upload an image to Discord. This can be in a direct message with the bot or in any channel.
Once uploaded, right-click the image (or long-press on mobile) and select Copy Link.
Paste this image link at the very beginning of your /imagine prompt. You can use multiple image links.
For example, to combine the style of a Van Gogh painting with a photograph of a city, your prompt might look like this:
Plain Text
/imagine https://vangogh-url.jpg https://city-photo-url.jpg a modern city skyline in the style of Starry Night
MidJourney will then attempt to blend the aesthetic elements of the reference images with your text prompt. This is an incredibly powerful technique for style transfer and creating unique hybrid visuals.
Controlling Randomness with Seeds
A seed number is a value that controls the initial field of visual noise from which your image is generated. By default, MidJourney uses a random seed for every prompt. However, by specifying a seed, you can create more consistent and reproducible results.
To find the seed of a previous generation, react to the job with the envelope emoji (✉️ ). The bot will send you the job details, including the seed number.
To use a specific seed, add the --seed <number> parameter to the end of your prompt.
Reusing the same seed with the same prompt will produce a nearly identical grid of images. This is invaluable when you want to make small tweaks to a prompt while keeping the overall composition the same. For example, you could keep the seed constant while changing only the lighting or color palette to see controlled variations.
Version 7 Superpowers: Unlocking Consistency
MidJourney Version 7 introduced game-changing features for maintaining consistency across multiple generations, addressing one of the biggest challenges in AI art.
Style Reference (--sref)
Style Reference is a powerful feature that allows you to apply the aesthetic of a reference image to your prompts with remarkable accuracy. Unlike a simple image prompt, --sref is specifically designed to extract and replicate the overall style—including color palette, lighting, composition, and texture.
To use it, simply add --sref <image_url> to the end of your prompt. This is the ultimate tool for maintaining a consistent brand aesthetic, creating a cohesive series of illustrations, or emulating a specific artistic style across different subjects.
Character Reference (--cref)
Character Reference is the solution to creating consistent characters across different scenes, poses, and outfits. By providing a reference image of a character and using the --cref <image_url> parameter, you can instruct MidJourney to maintain that character's facial features and general appearance in new generations.
You can further control the strength of the reference with the --cw (character weight) parameter, from --cw 0 (face only) to --cw 100 (face, hair, and clothes). This is an essential feature for storyboarding, creating comics, or any narrative project where character consistency is key.
Essential Parameters Cheat Sheet
Beyond the basics, a handful of parameters will give you a much greater degree of creative control:
--aspect or --ar: Sets the aspect ratio of your image (e.g., --ar 16:9 for widescreen, --ar 9:16 for vertical, --ar 1:1 for square).
--no: Negative prompting. This tells MidJourney what to exclude from the image (e.g., --no people, --no text).
--chaos <0-100>: Adjusts the randomness and variety of the initial image grid. A higher value creates more unexpected and diverse results, while a lower value produces more reliable, similar variations.
--stylize <0-1000>: Controls the strength of MidJourney's inherent artistic style. Lower values adhere more closely to your prompt, while higher values are more artistic and opinionated.
--tile: Generates images that can be used as repeating tiles to create seamless patterns for fabrics, wallpapers, and textures.
--weird <0-3000>: An experimental parameter that introduces quirky and unconventional aesthetics into your images, perfect for creating surreal and unexpected art.
From Beginner to Pro: The Path to Mastery
Mastering MidJourney is a journey of continuous experimentation and practice. By moving beyond simple prompts and embracing the full suite of advanced parameters and features, you can elevate your creations from interesting images to professional-grade works of art. The key is to develop a systematic workflow, understand how different parameters interact, and build a personal library of styles and prompts that work for you.
