As a creative, knowing how to craft Midjourney prompts is your ticket to faster, more consistent and unique designs. Get familiar with image generation best practices and parameters, and apply these concepts alongside me with example prompts and suggestions to improve them. You’ll be a Midjourney master in no time!
When you use AI to generate images for the first time, it’s a pretty magical experience.
But when you really dig in and realize how many benefits AI image generation tools like Midjourney offer, it’s game-changing. From faster sourcing to greater consistency and uniqueness, it’s no surprise Midjourney has gained so much traction in the past year.
But, just like adopting any new tool, it takes time to learn how to use AI for design.
So, in this article, I’ll walk through how creatives can craft the best Midjourney prompts in 2024 and lean on AI to scale design processes without compromising on quality.
Midjourney is built on Discord’s servers. That means you’ll need a Discord account to access and be able to interact with the Midjourney bot.
When opening up Discord, I’d recommend taking the direct message route when interacting with Midjourney bot. That way, you won’t cloud any public or group channels you may join.
To interact with the Midjourney bot, you need to first pick a command. Every command starts with a forward slash (“/”).
While there are many Midjourney commands you can pick from, the most common one is “/imagine.” This command prompts Midjourney to generate all sorts of images for you. But we’ll get into that more a little later on.
Some other commands include:
Comfortable with these basics and ready for more advanced commands? Check out AI expert Rory Flynn’s Ultimate Midjourney Power User Cheat Sheet. Then, keep reading to dig into Midjourney best practices, parameters and those examples I promised.
Infusing your creative vision into an AI tool, whether you use it to generate logos, presentations, or simply for art creation, doesn’t have to be an exhaustive task.
The generative AI tools market is expected to hit 137 billion US dollars by the end of 2024 (according to Statista), highlighting how learning to use Midjourney to generate amazing images with unique charm goes beyond a mere hobby in the creative journey.
And that’s the magical thing about Midjourney—when you input text into the prompt box, it can generate captivating visuals that align with your final image vision, but only if you do it right.
With the latest AI trends and developments in advanced prompts using Midjourney, it has become easier for creatives to master their text prompts until they achieve the desired outcome, whether it be beautiful images that resemble an oil painting, floating islands, or landscapes of natural beauty with vibrant colors.
At Superside, we are more than just an AI design agency. Our human creatives use AI tools with effective prompts to scale, differentiate, and reduce costs for our customers, and we have done so over 500 times in the past.
That’s where the text prompt becomes critically important, as it is the description that the stable diffusion model uses to generate the stunning visuals we envision.
For Ava Labs, our creatives used Midjourney to save over $1,000 and reduce design time by over 37%, generating more than 3,300 on-brand images for our customer.
From detailed character design to distinct artistic styles and everyday objects, we certainly know a thing or two about mastering Midjourney image prompts for consistency, on-brand images, and more (check out how we used Superside to achieve a long-sought goal in the AI world with epic detail and character consistency).
But AI isn't the end-all-be-all, and that’s why the process of creating prompts that match the intricate details of the illustration or art style you have in mind is so important to us.
There are a lot of similarities in how you’d write prompts for ChatGPT and Midjourney. Let’s walk through some best practices to consider when writing prompts for Midjourney.
Using ChatGPT for Midjourney prompts can significantly enhance the quality and creativity of the images generated. It’s a process that mixes the power of two AI tools to generate the image prompt you have in mind, allowing you to tweak the text description based on your needs, which is certainly of relative importance.
This is how the process would look like
1. Understanding Your Vision: Before you start, have a clear vision of what you want to create. Describe your idea in detail, considering elements like style, mood, colors, and specific features.
This is how the idea for the reference images would look like:
User: "I want to create a serene beach scene at sunset."
ChatGPT: "A tranquil beach at sunset, with golden hues reflecting on calm waves, silhouetted palm trees, and a clear sky transitioning from orange to deep blue."
2. Crafting Detailed Descriptions: Use ChatGPT to help you articulate detailed and vivid descriptions. For instance, based on the previous prompt, this is how refining detailed descriptions can look like:
User: "I want the waves to be more prominent and the sky to have a few scattered clouds."
ChatGPT: "A serene beach at sunset, with golden hues reflecting on prominent, gentle waves, silhouetted palm trees, and a clear sky with a few scattered, wispy clouds."
3. Incorporating Different Styles: Specify the art styles you want. Whether it’s a retrowave portrait, watercolor sketch, or digital art deco, ChatGPT can help you phrase your prompt to include these details, ensuring the generated image aligns with your aesthetic preferences.
4. Iterative Refinement: Use ChatGPT for iterative refinement of your prompts. Start with a basic prompt, generate an image, and then refine your description based on the results. ChatGPT can help you tweak and enhance your prompts to achieve the detailed illustration you are looking for.
5. Experimenting with Variations: ChatGPT can suggest variations of your prompt to explore different artistic possibilities. This helps in discovering unique and unexpected results that you might not have considered initially. Complementing the previous example, this is how the process would look like:
User: "How about adding a sailboat in the distance?"
ChatGPT: "A tranquil beach at sunset, with golden hues reflecting on gentle waves, silhouetted palm trees, a sailboat in the distance, and a clear sky with scattered clouds transitioning from orange to deep blue."
And finally, this is how this image would look like:
When writing your prompt, keep it simple. Single words (or even an emoji!) will produce an image, thanks to the default value of the tool. But, simple doesn’t always mean less.
If you’re looking for a more generic image, or don’t have a specific style you’re committed to, short prompts will rely on Midjourney’s default style. That means if you’re after a more unique look, a more descriptive prompt will produce a better result.
Clear and direct prompting equals clear and direct output. Like ambiguous prompting equals ambiguous output.
However, super long prompts aren’t always better either. Concentrate on the main concepts you want to create and exclude the rest.
Let’s look at three prompt examples and the images they generate:
Simple prompt: A dog.
Descriptive prompt: A dog chasing a frisbee while running in a field. Realistic. Wide angle shot.
Unnecessarily lengthy prompt: A Labradoodle dog chasing a frisbee while running in a field that’s surrounded by happy people. Make it look realistic. Generate the image as if it was taken with a wide angle lens. Use the brand color #5D3FD3 all over.
Sometimes, the words we use aren’t being understood in the way we want them to be. Just like when reading a creative brief, we might need to seek clarification around certain terminology. That clarification comes in the form of synonyms.
So, if you’re finding the AI output isn’t quite hitting your intended request, try different words.
Prompt: A student studying in a library.
Hypothetically, let’s say the generated image didn’t match what we wanted. Let’s try using synonyms like:
You get the idea.
When you feel like the generated images aren’t quite hitting the mark of your request, try mixing up your vocabulary.
Unlike humans, the Midjourney Bot doesn’t understand grammar or sentence structure. So, don’t worry about being grammatically perfect when writing prompts.
You can still use commas, brackets and hyphens to help organize your thoughts, but just know the Midjourney Bot might not reliably interpret them.
Think about blue elephants and don’t think about pink flamingos.
What animal are you picturing?
If the answer is pink flamingos, you may have more in common with AI than you might think!
AI works in a similar fashion to a child when you tell them not to push the red button. All they think about is pushing the red button, and in most cases, that’s the “output” that comes out.
So, when writing your prompts, avoid negative language at all costs. Let’s look at a couple examples of turning negative language into positive when writing prompts:
🚫[/imagine] an image of a cat. Avoid the color white.
✅[/imagine] an image of a brown cat
🚫[/imagine] a drawing of a busy street. Don’t make it look like a cartoon.
✅[/imagine] a drawing of a busy street. In the style of a sketch artist.
Think about creative briefs you’ve written or received in the past. Chances are you can remember a time when an important detail made the difference between hitting the mark or shooting an airball.
These types of details are equally important in your Midjourney prompts. Here are some examples:
For the sake of practicality, let’s walk through two key Midjourney parameters you’ll likely run into and need when creating AI-generated images: Aspect ratio and stylize.
Aspect ratio is the width-to-height ratio of an image. It’s typically expressed as two numbers separated by a colon, such as 7:4 or 4:3. This parameter changes the aspect ratio of your generated images, which comes in handy when you’re using the same graphic across Facebook, LinkedIn, a blog post and an Instagram ad.
You can specify one ratio in your original prompt and then continue to modify the ratio of your desired graphic post-generation. This useful parameter helps you scale your graphic design processes with AI.
To specify your desired aspect ratio in a prompt, simply add this at the end of your prompt “-- ar [add ratio here]”
It’s important to understand the Midjourney Bot has been trained to produce images that favor artistic color, composition and forms. That’s where the stylize parameter comes in handy. This parameter influences how strongly Midjourney’s training is applied.
Getting the style right is crucial because, when used correctly, Midjourney can bring your wildest visions to life. From illustrations for retro games to a neon-lit cityscape and cinematic lighting in geometric shapes, everything you envision can be brought to life if the artistic style is nailed in the prompt.
Wondering what all these best practices and commands look like in action? Let’s walk through some examples of the cool ways you can leverage AI in the design creation process with tips to refine each Midjourney prompt further.
I’ve broken out the following Midjourney prompts and designs by style.
If your team is into Star Wars like me, I think you’ll enjoy the theme…
Prompt: Darth Vader in the toilet eating noodles --s 750
Output:
Suggestions for prompt modifications:
Prompt: A panda bear with cyberpunk, hyper-realistic, technological accessories, blue and fuchsia lighting around in an alley of the future. The panda looks very cool and has sunglasses, the panda is strong but not fat, holograms coming out of one of his mechanical arms
Output:
Suggestions for prompt modifications:
Prompt: futuristic cityscape with a car, accelerating on a road in the middle of a city full of neon signs, the camera lens is focusing the car
Output:
Suggestions for prompt modifications:
Prompt: a breakfast inside a glass cube with elements floating on a 3d pop aesthetic with a pink background
Output:
Suggestions for prompt modifications:
Prompt: A stormtrooper’s selfie, the stormtrooper is cooking with yellow ingredients, is in the middle of a kitchen.
Output:
Prompt improvement suggestions:
Prompt: handsome charming Jesus taking a selfie with his apostles, wide angle shot, happy and smiling, hands and fingers with positive sign for the camera, Jesus has long hair and beard, Jesus has blue eyes, jubilant, vivacious, they’re in a cave with a big rock as a door opened, sunrise lighting --ar 32:18 --v 5 --q 2 --s 750
Output:
Prompt improvement suggestions:
AI will never replace designers, or marketers for that matter. That’s a hill I’m willing to die on. But, that doesn’t mean AI won’t transform the creative industry and become embedded into existing creative workflows to help speed up tasks, fuel innovation and expand capacity.
That’s why it’s so important to start building your Midjourney muscles as a designer—so you can learn how to best embed AI into your existing design processes and start reaping the benefits for your brand.
Hiba Amin is a Contributing Writer at Superside. As a marketing leader who lives and breathes content, she's had the privilege of heading up content teams and has also been in the trenches as a marketing team of one. She's worked at a wide range of tech companies across PLG, SaaS and most recently, AI. Say hi to her on LinkedIn (and ask her about her dog, Milo!).
Get high-quality creative, ship campaigns faster and stand out from the competition.
Never say no to another project request. Get a hassle-free creative partner that can keep up.
Allow your in-house creatives to focus on more strategic projects. Get new ideas & continuous design inspiration.
Increase your design capacity without additional hiring and with fewer vendors to manage.
Get a demo and discover how 450+ ambitious companies and 2,500 energized fans use Superside to free themselves from the shackles of limited budgets, broken processes and stretched in-house teams.