Start Making AI Art: A Simple Guide to Using Midjourney

Ready to turn your ideas into images? Midjourney is a powerful tool that lets you do just that, using simple text commands. This guide cuts through the confusion and shows you how to start creating stunning AI art, from your very first image to more controlled and refined results.

Whether you're new to AI art or looking to get more consistent outputs, this post covers the basics, how to add detail with different settings, and some advanced tips to help you get the images you want. We'll explore how Midjourney understands your words and how you can guide it to create specific styles, scenes, and effects.

Getting Started with Midjourney

Midjourney works through Discord, a free chat application. If you don't have it, download Discord from discord.com and create an account.

[P]Once Discord is set up, you need to add the Midjourney bot to your own server.

[/UL][/P]

Midjourney requires a paid subscription. You can start the subscription process directly in Discord using the `/subscribe` command. This gives you a link to choose a plan based on your needs and make payment.

Creating Your First Image

The core command in Midjourney is `/imagine`. Type this followed by a space, and then enter your prompt – a text description of the image you want to create.

Midjourney takes your words and phrases, called tokens, and compares them to its training data to generate images. Simple prompts work fine, but the fewer words you use, the more weight each word carries. You don't need to add phrases like "make me an image." Just describe what you want.

Understanding the Image Grid

When you use `/imagine`, Midjourney gives you a grid of four low-resolution images. These images are arranged in a Z-pattern: top-left is 1, top-right is 2, bottom-left is 3, and bottom-right is 4.

• U Buttons (U1, U2, U3, U4): These let you upscale the corresponding image in the grid to a higher resolution. Choose the one you like best.

• V Buttons (V1, V2, V3, V4): These create new variations based on the selected image. Midjourney uses the chosen image as inspiration to generate four similar but different options.

• Regenerate Button: The circular arrow lets you run the exact same prompt again to get a new set of four images.

Refining Your Upscaled Image

Once you upscale an image, more options appear.

• Vary (Strong / Subtle): Generate variations that are either quite different (Strong) or only slightly different (Subtle) from the upscaled image.

• Vary (Region): Select a specific area of the image and write a prompt to change only that part.

• Pan Buttons (Up, Down, Left, Right): Extend the image scene in a specific direction. This is like moving your camera to show more of the surrounding area.

• Web Button: Opens the image in your Midjourney web gallery.

• Favorite Button: Marks the image as a favorite in your web gallery.

Adding Detail with Prompt Variables

To get more specific results, you can add descriptive keywords often called variables. These help guide Midjourney on the look and composition of your image.

• Artistic or General Style: Specify art movements (e.g., Pop Art), artistic styles (e.g., Pixar cartoon style), or visual vibes (e.g., editorial style).

• Composition: Define the camera shot (e.g., wide shot, full body shot, close-up, aerial shot, low angle shot).

• Medium: Indicate how the image should look like it was made (e.g., sketch, oil painting, watercolor, film still).

• Camera/Film Type: Suggest the look of specific cameras or film (e.g., Kodak, Canon, Fujifilm). You can even mention lenses or effects like bokeh.

• Subject Description: The basic idea, but adding specific details about clothing, expressions, or actions.

• Environment: Define the location or setting (e.g., city street, garden, war zone, fantasy land).

• Lighting: Specify the type of light (e.g., natural sunlight, studio lighting, cinematic lighting, neon lighting, side lighting).

• Atmosphere: Describe the mood or weather (e.g., sunny, rainy, misty, foggy).

• Mood: Suggest the feeling or tone of the image (e.g., romantic, dreamy, luxurious).

• Rendering: Use terms like photo realistic, 4K, 8K, hyper realistic, or mention rendering engines like Unreal Engine or Octane Render to influence detail and quality.

By combining a basic prompt with these variables, you get much closer to your desired image. Experiment with different combinations and see what works for you.

Managing many prompts and variations can become time-consuming. If you find yourself needing to generate images faster or in bulk, exploring tools designed for workflow can help. The TitanXT Midjourney Automation Suite offers features to streamline your image creation process.

Advanced Midjourney Parameters

Once you're comfortable with the basics and variables, you can use parameters to gain even more control.

Negative Prompting (--no)

Use `--no` followed by keywords you want Midjourney to try and avoid in the image. For example, `--no trees, plants` tells Midjourney not to include trees or plants.

Prompt Weights (::)

Add `::` followed by a number after a word or phrase to give it more or less importance. `futuristic::2 metro` gives the word "futuristic" twice as much weight as "metro." All parts of a prompt have a default weight of 1 if not specified.

Multi-Prompt (::)

Using `::` also helps Midjourney see different parts of your prompt as separate ideas rather than one blended concept. `cricket:: bat` treats "cricket" and "bat" as distinct elements, while `cricket bat` might be seen as a single concept.

Aspect Ratios (--ar)

Control the shape of your image grid or upscaled image using `--ar width:height`. Examples include `--ar 16:9` for landscape or `--ar 9:16` for a vertical image. This is key for things like full-body shots or specific social media formats.

Seed (--seed)

Midjourney uses a random seed number for each image generation. Using the same prompt with different seeds gives different results. If you find an image style you like, you can get its seed number (by reacting with an envelope emoji on the image in Discord) and use the `--seed` parameter with that number in future prompts. This helps create images with a similar visual style.

Stylize (--s)

The `--s` or `--stylize` parameter controls how much of Midjourney's default artistic styling is applied. Values range from 0 (very little styling, closer to your exact prompt) to 1000 (very strong styling, more artistic and surprising results). The default is 100.

Chaos (--c)

The `--c` or `--chaos` parameter affects how varied the four images in the initial grid are from each other. A low value (0-20) means the images will be quite similar. A high value (50-100) leads to wildly different results.

Repeat (--repeat)

The `--repeat` parameter lets you run the same prompt multiple times automatically. `--repeat 5` runs the prompt five times, giving you 20 images (5 sets of 4). This is useful when you want to quickly generate many variations of a concept.

Incorporating multiple parameters can speed up finding the perfect image. Automation tools can further assist by managing complex prompt structures and repeat jobs efficiently. Consider using the TitanXT Midjourney Automation Suite to handle these advanced techniques more easily.

Cool Hacks and Useful Resources

Scaling Images Further

While Midjourney upscales images, external tools can increase resolution even more for high-detail needs. Websites like Upscale.media or Vectorizer.ai (for turning logos/simple art into scalable SVG files) can be helpful for post-processing your images.

Face Swapping

There are Discord bots like 'InsightFaceSwap' that allow you to swap faces in generated images. Install the bot, save a face ID using `/saveid`, set your desired face using `/setid`, and then use the 'Inswapper' app on a Midjourney image to swap the face.

Blending Images

The `/blend` command lets you upload 2 to 5 images, and Midjourney will combine their concepts and aesthetics into a new, unique image.

Describe an Image

Use the `/describe` command. Upload an image, and Midjourney will provide four sample prompts it thinks could generate a similar image. This is a great way to learn how to phrase prompts for specific visual styles.

Using Images as Prompts

You can start your `/imagine` prompt with an image link. Midjourney will then use that image's style or composition as inspiration for the prompt that follows it. Upload an image to Discord, get its link, and paste it at the start of your prompt.

Creating Anime Style Images

Add the parameter `--niji` to your prompt. This activates Midjourney's anime-specific model, tailored for creating illustrations with an anime aesthetic. You can combine this with other style and composition parameters.

Resources for Inspiration and Prompt Building

• Midjourney Website (Explore): See images created by others and copy their prompts to learn.

• Prompt Hero: A large collection of prompts for various AI art tools, categorized by style and subject.

• The New Computer Science / WikiArt: Websites offering insights into different art styles and movements you can use as keywords.

• Community-Made Prompt Builders: Spreadsheets or websites designed to help you select keywords and parameters to build complex prompts without memorizing everything.

Making Midjourney Work for You

Beyond personal experimentation, Midjourney has many commercial use cases. You can use it for creating assets for games, illustrations, marketing materials, product mockups, website elements, and much more.

Consistency in style for projects like games or branding often relies on using parameters like `--seed` and carefully defining styles. Post-editing in software like Photoshop is also key to fixing details and getting images production-ready.

Midjourney is a tool for creative exploration. Don't be afraid to try new things, combine different parameters, and see what surprising results you get. The more you experiment, the better you'll understand how to achieve your vision.

For creative professionals and businesses, managing a high volume of Midjourney images and ensuring consistent styles across projects can be a challenge. Tools like the TitanXT Midjourney Automation Suite provide features to help organize, automate, and scale your AI image generation workflows, making it easier to integrate AI art into your process.

Conclusion

You now have a solid foundation for using Midjourney. Start with simple prompts, gradually add variables to guide the style, composition, and feel of your images, and explore advanced parameters for fine-tuned control. Use the available resources to find inspiration and build better prompts.

AI art is a new creative frontier. Have fun exploring, experimenting, and seeing where your imagination takes you with Midjourney.