Next-Level Midjourney

Midjourney's text-to-image AI is an innovative tool that supercharges your creativity by leveraging AI and machine learning to craft stunning visuals from simple descriptions. In my previous article, I introduced you to Midjourney and walked you through your first image creations. Since then, Midjourney has evolved significantly, now at version 6. If you missed the earlier guide, you can catch up here to kickstart your journey with Midjourney.

In the previous article, we briefly touched on the basics of commands, parameters, and features to enhance your image generation experience. In this article, we'll dive deeper into these elements through practical examples and demonstrations, empowering you to unleash your imagination like never before. We'll zoom in on the most crucial commands, features, and parameters for image generation. Once we've covered these essentials, we'll delve into the revamped prompt structures introduced in version 6.

Today's content includes:

• Commands

• Parameters

• Features

• Prompt Structures

Let's kick things off by exploring the commands in detail!

Commands

To begin creating your images, users must engage with the Midjourney bot on Discord by issuing “commands”. These commands serve as versatile tools, enabling users to interact with Midjourney in a manner aligned with their creative objectives and preferences. These commands start off with a ‘/’ followed by the command you want. I will brake these into commands down into four distinct groups: Creation, Efficiency, Privacy, and Account commands. As stated before, we will be focusing on more prominent commands that are utilized in the creation process. So, we will not go into Account commands since those pertain particularly to account information and settings.

Creation Commands

These commands enable users to both create and blend images, while also aiding in the formulation and condensing of prompts for Midjourney to enhance its image generation capabilities. In essence, they play a vital role in the creative process. Below, you'll find a list of commands commonly used for image generation:

• /imagine: The primary command you will utilize the most for generating images. Produces a set of four images based on a given prompt as shown in Figure 1.1:

Figure 1.1 - Prompt: a gleeful smiling puffy cartoon cloud --v 6.0

• /blend: With this command, you can upload anywhere from two to five images, which will then be seamlessly integrated together. Observe the result of the three images blended into the single image in Figure 1.2:

Figure 1.2 - Using the /blend command combined the three images into a single image.

• /describe: Using this command, Midjourney will take a single uploaded image and generate four text prompts that describe it. Observe Figures 1.3 and 1.4 to see the results:

Figure 1.3 - Using the /describe following by the image you upload will generate four prompts based off the image.

Figure 1.4 - Choosing option 4, it will generate a new set of images based off that prompt.

• /shorten: If you find yourself struggling to condense lengthy prompts, this command is designed for you. It'll generate five shortened variations of your prompt and provide guidance on how to streamline the content for greater conciseness. Observe Figure 1.5 below:

Figure 1.5 - Using the previous example’s prompt, we can /shorten command to concise it down to five shorter options.

Figure 1.6 - Choosing option 1, it will generate a new set of images based off that prompt.

Efficiency Commands

These commands provide you with the flexibility to adjust the speed at which your images are generated. It's worth noting that only Standard, Pro, and Mega subscription plans can take advantage of the relax hours, and each plan differs in the allocation of fast hours. You can change your subscription by accessing your account using “/info” command and logging onto Midjourney website. Below are the commands that enable you to tailor the speed of your image generation:

• /fast: The default setting when you start Midjourney and the only option for Basic plan users. This command activates fast mode, accelerating the image generation process for improved efficiency.

• /relax: Exclusive to Standard, Pro, and Mega plans, this command switches to relax mode, optimizing image generation for a more deliberate and controlled pace. Plans with this feature enjoy unlimited generation hours, albeit with longer wait times for image generation, depending on the workload of Midjourney's GPUs.

• /turbo: This command switches to turbo mode to significantly speed up image generation, surpassing the pace of the /fast mode. This advanced setting operates four times faster, albeit consuming twice the GPU minutes. Turbo mode is exclusively available in Midjourney version 5 and beyond.

Privacy Commands

These commands offer you the ability to customize settings regarding the visibility of your generated images to other users on both the Midjourney website and the Discord server. Please note that not all Midjourney users will have access to these privacy features, depending on their subscription plan. Below, you'll find some privacy commands to adjust the visibility of your images:

• /public: This serves as the default setting when you begin using Midjourney. This command allows other users on both the Midjourney website and Discord server to see your generated images.

• /stealth: With this command, you can keep your generated images private and confidential, ensuring they don't appear on the Midjourney website. This exclusive feature is available to subscribers of the Pro or Mega plans.

• /private: This command functions similarly to the /stealth command, but with the added benefit of keeping your generated images private and confidential, ensuring they won't appear on the Midjourney Discord server. Once more, this exclusive feature is available to subscribers of the Pro or Mega plans.

Parameters

Beyond commands, Midjourney provides customizable parameters essential for tailoring the image generation process to your creative vision. These parameters give you the ability to finely adjust different aspects of the generation, such as aspect ratio, rendering quality, and style. These parameters are added at the end of your prompt starting with ‘--’ followed by the parameter you want to change. Once again, I will be going over the parameters that I believe are most important for the creation process. Below is a compilation of commands that can be added to the end of your prompts to effectively manipulate the resulting image:

• Aspect Ratio(--aspect or --ar <value>:<value>): The aspect ratio parameter is a versatile tool for adjusting the dimensions of your AI-generated images. It enables you to modify the width-to-height ratio, typically represented as two numbers separated by a colon, such as 7:4 or 4:3. While the default aspect ratio in Midjourney is 1:1, you can easily alter it using whole numbers to suit your preferences.

Figure 2.1 - Prompt(Left): a desert wasteland experiencing a solar eclipse in the art style of tite kubo --niji 6 --ar 16:9;

Prompt(Right): a desert wasteland experiencing a solar eclipse in the art style of tite kubo --niji 6 --ar 9:16

• Chaos(--chaos or --c <0-1000>): This parameter in Midjourney is a fascinating tool used to shape the diversity of initial image grids. It adds an element of surprise and creativity to the image generation process, resulting in unexpected and artistic outcomes. When opting for higher chaos values, you'll encounter more unusual and unpredictable compositions. Conversely, lower values provide more reliable and consistent images. By default, the chaos parameter is set to 0 and accepts values from 0 to 100. If you don't specify a chaos value or use a low one, the original grid image will remain similar each time you generate images. However, as you increase the chaos value, the original grid image becomes more varied, leading to unexpected creations. To adjust the chaos value, simply append '--chaos <value>' or '--c <value>' to your prompt. This parameter significantly influences the variation in generated results, with higher values, such as '--chaos 100', resulting in more unpredictable and unusual creations, while lower values, like '--chaos 0', produce more consistent outcomes.

Figure 2.2 - Prompt(Left): a close up face profile of a goth woman in an abstract background with vibrant colors of blues, purples, and pinks --c 10;

Prompt(Middle): a close up face profile of a goth woman in an abstract background with vibrant colors of blues, purples, and pinks --c 50;

Prompt(Right): a close up face profile of a goth woman in an abstract background with vibrant colors of blues, purples, and pinks --c 90

• Character Reference(--cref [URL] + optional: --cw <1-100>): The character reference parameter that was recently introduced in version 6 that helps you keep your characters consistent across different images and styles. Before Character Reference existed, it was tricky to generate more images of the same character. You could try using image prompts, but they often didn't capture the character's essence in various poses. With Character Reference, you can create a character based on an existing image of your character. Just add --cref followed by a URL of your character's image to your prompt. If you want to adjust how much the reference image influences your new image, you can use character weight or --cw from 0 to 100.  Higher character weight considers more of the face, hair, and clothes of your character, while lower character weight particularly focuses more on just the face. The character reference parameter works best with images generated by Midjourney that feature a single character or fictional characters.

Figure 2.3 - Reference(Top);

Prompt(Left): full length, character reference sheet, multiple poses, space warrior, mecha suit, helmet, arm cannon, bulky shoulder armor, orange visor, steampunk --v 6.0 --cref [URL] --cw 10;

Prompt(Middle): full length, character reference sheet, multiple poses, space warrior, mecha suit, helmet, arm cannon, bulky shoulder armor, orange visor, steampunk --v 6.0 --cref [URL] --cw 50;

Prompt(Right): full length, character reference sheet, multiple poses, space warrior, mecha suit, helmet, arm cannon, bulky shoulder armor, orange visor, steampunk --v 6.0 --cref [URL] --cw 100

• Image weight(--iw <0-2>): This parameter determines the relative importance of the image in comparison to the text prompt. Values exceeding 1 signify a higher priority on the image, whereas values below 1 prioritize the text prompt more. The default setting for image weight is 1.

Figure 2.4 - Reference(Top);

Prompt(Left): [URL] space warrior, white and blue mecha suit, blue helmet, red arm cannon, blue bulky shoulder armor, red visor, steampunk, running --v 6.0 --iw .5;

Prompt(Right): [URL] space warrior, white and blue mecha suit, blue helmet, red arm cannon, blue bulky shoulder armor, red visor, steampunk, running --v 6.0 --iw 2

• Niji(--niji or --style<original, cute, expressive, scenic>): This parameter opens up the possibility to utilize an alternative model tailored specifically for generating anime-style images. Version 5 and on, the Niji Model offers flexibility through ‘--style"’ parameters, enabling users to achieve distinctive appearances. These Niji Style Parameters encompass various options: ‘--style cute’ generates endearing characters, props, and settings with a charming appeal, while ‘--style expressive’ brings forth a more advanced illustrated feeling. ‘--style original’ maintains the essence of the original Niji Model Version 5. Additionally, ‘--style scenic’ focuses on crafting breathtaking backgrounds and cinematic character moments within fantastical surroundings, enriching the visual narrative experience.

Figure 2.5 - Prompt: full length, character reference sheet, multiple poses, pretty goth woman, black hair brindled blue, nose piercing, arm tattos, black metal band shirt, black jeans, black belt with skull buckle, chain hanging off belt, black converse shoes, the art syle of Kōhei Horikoshi --niji 6

• No(--no): Known as negative prompting, the 'no' parameter in Midjourney is a useful tool for specifying elements you want to exclude from your generated image. This feature is pivotal in Midjourney prompts, giving you the ability to filter out unwanted elements from the final images. You can easily apply this parameter alongside multiple words or parameters, separated by commas, like '--no item1, item2, item3, item4'. For instance, let's say you're creating a still-life painting, but you prefer not to include any fruit. In that case, you would simply use '--no fruit'. Observe Figure 2.6 and notice how the left set of images still generated some fruit in them without the word being in the original prompt:

Figure 2.6 - Prompt(Left): a still life painting of a basket --v 6.0; Prompt(Right): a still life painting of a basket --v 6.0 --no fruit

• Stop(--stop <1-100>): The '--stop' parameter in Midjourney gives you the ability to end a job before it's completed. This feature allows you to halt the image generation process at a specific percentage set within your prompt. The stop parameter accepts values ranging from 10 to 100, with the default being 100. For example, using the '--stop 75' command directs Midjourney to stop generating the image once it reaches 75% completion. This can be particularly useful for achieving a Gaussian blur effect in your images. It's worth noting that the stop parameter doesn't function when upscaling.

Figure 2.7 - Prompt(Left): A fruit tree full of apples, bananas, watermelons, oranges, grapes, pineapples, cherries, blueberries, peaches, and lemons. --v 6.0 --stop 100;

Prompt(Right): A fruit tree full of apples, bananas, watermelons, oranges, grapes, pineapples, cherries, blueberries, peaches, and lemons. --v 6.0 --stop 60

• Style Reference(--sref [URL] + optional: --sw <0-1000> and --sv <1-4>): The style reference parameter functions similarly to the character reference parameter works. The Style Reference was also introduced in version 6 of Midjourney and lets you borrow the style of a reference image by simply adding its URL after the ‘--sref’ parameter. This tool is super handy for designers sticking to a brand's look or for creatives working on projects that need a consistent visual vibe. Simply just add --sref followed by a URL of your image with the style to your prompt. The style reference parameter comes with four variations that influence the style of your image. You can use these variations with the ‘--sv’ followed by a number from 1 to 4:

1. --sv 1 uses the original style reference algorithm, more “vibey”

2. --sv 2 uses the second iteration of style references

3. --sv 3 uses an updated version of the more “vibey” --sv 1

4. --sv 4 uses an updated version of --sv 2

The style reference parameter also uses a style weight parameter --sw to set the strength of the stylization ranging from <0-1000>. A higher number will make the image follow the reference image more closely while a lower number will ignore most of the style of that reference image.

Figure 2.8 - Character Reference(Left); Style Reference(Middle);

Prompt(Right) - Samus Aran from Metroid in the art style of Tite Kubo --niji 6 --cref [URL] --cw 100 --sref [URL] --sw 2

• Stylize(--stylize or --s <0-1000>): The '--stylize' or '--s' parameter in Midjourney influences the degree to which the bot's training is applied, resulting in images with distinct artistic qualities in terms of color, composition, and forms. This parameter directly impacts how much of Midjourney's training is reflected in your images. By default, the stylize parameter is set to 100, and it accepts integer values from 0 to 1000. To adjust the stylization level, simply append '--stylize <value>' or '--s <value>' to your prompt. Lower values tend to produce images closely aligned with the prompt but may exhibit less artistic flair. Conversely, higher values generate highly artistic images, albeit with a decreased connection to the original prompt. It's important to note that different versions of Midjourney and model versions offer varying stylize ranges. For example, versions 5, 5.1, and 5.2 feature a stylize range of 0 to 1000.

Figure 2.9 - Prompt(Left): a cyberpunk metropolis floating above rainy clouds with vibrant colors of red, pink, purple, and blue --v 6.0 --s 200;

Prompt(Right): a cyberpunk metropolis floating above rainy clouds with vibrant colors of red, pink, purple, and blue --v 6.0 --s 800

• Tile (--tile): The tile parameter in Midjourney helps create seamless patterns. It takes your prompts and turns them into images that can be repeating patterns. These images are great for designing fabrics, wallpapers, and textures, giving your projects a polished and cohesive look.

Figure 2.10 - Prompt: apples, bananas, oranges, watermelons, grapes, lemons, pineapples, cherries, blueberries, and peaches --v 6.0 --tile

• Version(--version or --v <1, 2, 3, 4, 5, 5., 5.2, or 6>): The '--version' or '--v' command provides you with the flexibility to choose from a range of iterations of the Midjourney algorithm. While the latest model version serves as the default choice, specifying a version number allows you to opt for the iteration best suited to generate your preferred image type. As a leading platform in AI-generated imagery, Midjourney is committed to continuous innovation and improvement in model versions. The platform regularly introduces new models to enhance performance and output, offering users a diverse array of aesthetic possibilities. Version 5.2 is tailored to deliver more detailed and sharper results, featuring enhanced colors, contrast, and compositions. Notably, this model demonstrates improved understanding of prompts and is highly responsive to the full range of the '--stylize' parameter. Version 5.1 streamlines usability with simple text prompts and excels in accurately interpreting natural language prompts. It produces fewer unwanted artifacts and borders, while also supporting advanced features like repeating patterns with the '--tile' parameter. Version 5 closely mirrors the prompt but may require longer prompts to achieve the desired aesthetic. It's an ideal choice for users seeking a more realistic output. Version 4, developed by Midjourney and trained on the new Midjourney AI supercluster, is particularly adept with image prompts, boasting high coherency and exceptional performance.

Figure 2.11 - Prompt(Left): an old man at the park sitting on a park bench overlooking a lake --v 6.0;

Prompt(Right): an old man at the park sitting on a park bench overlooking a lake --v 3

Extra Parameters

• Fast, Slow, or Turbo(--fast, --slow, or --turbo): Using the '--fast', '--relax', or '--turbo' parameter in your prompt allows you to specify the GPU speed for your job. However, it's important to note that if you include this parameter at the end of a prompt, it will only run that single job using the specified GPU speed, regardless of your preferred settings. For instance, if your preferred setting is 'fast' through the Midjourney Commands, but you include the 'relax' parameter in your prompt, it will override the 'fast' setting and use the 'relax' parameter from your prompt instead.

• Quality(--quality or --q <.25, .5, or 1>): The '--quality' or '--q' parameter in Midjourney has a significant impact on the time it takes to generate an image. This parameter determines the duration allocated to the rendering process for image generation. For example, lower-quality settings might be ideal for abstract or landscape images, or for generating anime-style images. Conversely, higher-quality values may enhance the appearance of architectural images that benefit from intricate details. However, it's worth noting that the quality setting doesn't affect the resolution of the generated image. By default, the quality value is set to 1, accepting only values of 0.25, 0.5, and 1. Any values larger than 1 are automatically rounded down to 1. It's important to understand that higher values aren't always better. Depending on the image you're creating, lower-quality settings can sometimes yield superior results.

• Repeat(--repeat or --r <1-40>): With the '--repeat' or '--r' parameter, you have the ability to run a job multiple times using a single prompt. This feature allows users to specify the desired number of iterations for Midjourney to generate outputs based on the same prompt. When combined with parameters like '--chaos', '--repeat' accelerates your visual exploration journey. It's worth noting that '--repeat' is only available in Fast and Turbo GPU modes. Additionally, if you use the redo or re-roll button on the result of a '--repeat' job, it will only rerun the prompt once.

• Seed(--seed <0-4294967295>): The '--seed' parameter in Midjourney serves as a way to initiate an initial grid from a field of visual noise, providing a starting point for generating the initial image grids. While the Midjourney bot generates seed numbers randomly for each image prompt, you have the option to specify a seed number using the seed parameter. To adjust seed numbers, simply append '--seed <value>' to the end of your prompt. The seed parameter accepts whole numbers from 0 to 4294967295. Using the same seed number and prompt will result in very similar generated images and similar ending images. However, if you omit specifying a seed, Midjourney will utilize a randomly generated seed number, resulting in a diverse array of options each time a prompt is used. To obtain the seed number of a previously generated image, simply reply with an ✉️ emoji on that image.

Features

You may notice the rows of buttons beneath your generated images, labeled U1, U2, U3, U4, and V1, V2, V3, V4, along with a refresh button. Also, notice how each image is numbered, corresponding with these buttons. For instance, if you're drawn to the second image at the top right of the grid, simply select either U2 or V2 to focus on that particular image. Observe figure 3.1 below:

Figure 3.1 - The buttons in the red square are some of Midjourney’s features. The numbers on each image represent to the corresponding “U” or “V”. If you select “U1”, you will get a single image of the top left image by itself.

In earlier versions of Midjourney, the 'U' buttons were primarily used for upscaling images. However, in the current version, these 'U' buttons serve a different purpose—they allow users to select and detach an image from the grid. This not only streamlines the downloading process but also enhances editing and generation options.

Conversely, the 'V' buttons are dedicated to generating image variations. When clicked, they produce a new grid containing four distinct images that maintain the overall style and composition of the selected variation.

Additionally, the blue refresh button performs a similar function by re-executing the task, resulting in another set of four images.

Now, picture this: you've stumbled upon an image that catches your eye and have singled it out from the rest using whichever ‘U’ corresponds with that image. But perhaps you'd like to tweak or refine it further to suit your taste. Right below this individual image, you'll uncover a variety of options. Observe Figure 3.2 below:

Figure 3.2 - Here are more of Midjourney features you can utilize when working with a single image of your interest.

• Upscale(Subtle): This option doubles the size of your new image while staying similar to the original image.

• Upscale(Creative): This option also doubles the size of your new images but it will add new details that are different from the original image.

Figure 3.3 - Side by side comparison of Subtle Upscale(Left) and Creative Upscale(Right). Notice the background has more stars in the image than the right.

• Vary(Subtle): This button generates a new grid of four images with slight adjustments from the selected image.

• Vary(Strong): This button generates a brand new grid of four images, each incorporating significant and noteworthy alterations based on the chosen image.

Figure 3.4 - Side by side comparison of Subtle Vary(Left) and Strong Vary(Right). Notice the set of images to the right are posed different and have a different background.

• Vary(Regional): This button allows for precise selection of specific areas within the image for modification, while keeping the rest of the image intact. This feature comes in handy, particularly in cases where Midjourney may face difficulties with intricate details like hands and feet. Once you've made your selections, this option will generate a new grid of four images, concentrating on the specified regions that require modification from the original image.

Figure 3.5 - The checkered region represents the area where you want to make changes at. After, prompt you would like the changes to be made.

Figure 3.6 - Based on the region you selected for changes and the prompt you typed, Midjourney will generate four new variations making changes to the selected areas based off the new prompt.

• Zoom Out 1.5x, 2x, and Custom Zoom: These features expand the visual scope of the image while preserving its content. The widened perspective is intelligently filled using cues from the original prompt and image. With the 'Custom Zoom' option, you have the flexibility to define a specific aspect ratio beyond the standard 1.5x or 2x options.

Prompt 3.7 - Using 2x Zoom Out on the original image to the left, it generated more background scenery and some more fruits in the foreground in the image to the right.

• Left, Right, Up, and Down Arrows: Referred to as the pan buttons, these allow users to extend the image in a direction of their choice without changing the original content. The expanded pan area is cleverly populated using cues from the original prompt and image.

Prompt 3.8 - Similar to how Zoom Out functions, Pan options generate more scenery. However, instead of an image generation all around the original image, it generate more based on the direction you pan. This will also manipulate the aspect ratio of the image.

• Remaster: This lets you overhaul earlier, upscaled Midjourney images from previous versions to take advantage of newer rendering models. Whenever you use Remaster, it generates a new grid of images based on the original composition and combines it with the coherency of the newest Midjourney Model Version. Be prepared for substantial differences however — while the overall layout may look similar, expect colors and details to shift, possibly even the overall style. You may or may not find the output matches your intent, much like any original Midjourney prompt.

Figure 3.9 - If you have generated and singled out any images using version commands before version 5, you will see the Remaster feature.

Figure 3.10 - Using the Remaster feature, you can generate a new images that have more modern aesthetic to the latest version of Midjourney.

Prompt Structure Revisited

If you remember from my previous article, I discussed and introduced a valuable guide to help you structure your prompts as well as some fundamentals. Here is the prompt structure previously discussed:

[content type], [description + subject + adjectives], [style], [parameters]

I still recommend following this structure when creating prompts as well as making sure it is clear and concise while gradually introduce complexity, additional details, and styles. However, with the release of Midjourney version 6, several enhancements were introduced to help you create incredibly accurate and stunning AI-generated artwork from your prompts. Some of the key changes in Midjourney version 6 include reduced sensitivity to filler words, the ability to process more natural and conversational prompts, improved coherence in interpreting prompts featuring multiple subjects, actions, and settings, enhanced text generation capabilities, and the capability to understand longer prompts of up to 350 words. While these updates enhance and expand creative possibilities, they also offer more flexibility in prompting. Whether you're using the latest version of Midjourney, an older iteration, or the Niji model, the above structure still offers a solid foundation for prompt engineering.

However, in some of the previous prompt examples you may have noticed that they don’t all follow the above prompt structure and may require advanced prompts. Advanced prompts are a bit more complex than your standard prompts as they can include using one or more image URLs, multiple phrases of text, or permutation. I will quickly go over each of these methods and demonstrate how to use them.

Image Prompts

[Image URL], [description + subject + adjectives], [style], [parameters]

Integrating image URLs into your prompts can significantly shape and enhance the style and content of the generated result. Typically, image URLs should be placed at the beginning of your prompt as shown in the image weight example. However, if you recall the character and style reference examples, you add the URLs following those parameters instead of the beginning of the prompt. To include images, simply type or paste the web address where the image is hosted online. Ensure that the address concludes with a file extension such as .png, .gif, or .jpg.

Once you've added the image URLs, proceed to incorporate any additional text and parameters to refine and personalize your prompt. This approach allows for a seamless integration of visual cues, fostering a more nuanced and tailored output that aligns with your creative vision.

Multiple Prompts

Imagine you’re crafting a creative prompt for Midjourney, and you want to blend a couple of ideas. That’s where the magic of the :: separator comes in! It's like giving directions to Midjourney on how to mix things up.

By using :: in your prompt, you're telling Midjourney to pay attention to each part separately. For instance, if you type "cold heart," Midjourney creates a frozen heart. But if you write "cold:: heart," it gets interesting! Now, Midjourney treats "cold" and "heart" as distinct concepts, creates them as separate entities.

And here’s the kicker: you can tweak the importance of each part. Just add a number after :: to let Midjourney know which part matters more. So, "cold::2 heart" means cold takes the spotlight, making your imagery all about the freezing cold adventures while heart playing a supporting role. Observe Figure 4.1 below:

Figure 4.1 - Prompt(Left): cold heart; Prompt(Middle): cold:: heart; Prompt(Right): cold::2 heart

Permutation Prompts

This innovative tool allows you to generate multiple image variations from a single prompt, turbocharging your content creation efforts.

Here's how it works: just enclose a list of options within curly braces {} in your prompt, separated by commas. Each option becomes a separate "job" that Midjourney processes. For example, using the prompt "/imagine a {red, blue, yellow} car" creates three different images of cars in red, blue, and yellow.

Depending on your subscription tier, you can create a maximum number of jobs with a single Permutation Prompt. Basic subscribers get 4 jobs, Standard subscribers get 10, and Pro/Mega subscribers get 40.

Permutations are like playing mix and match with colors and objects, giving you endless creative possibilities. They streamline your workflow by quickly generating variations without manual tweaking. Plus, they accelerate innovation by enabling rapid testing and iteration, helping you find the perfect visual expression faster.

Figure 4.2 - Using a permutation prompt, you can generate a multiple variations of a prompt whether it’s for an adjective, a person, an object, etc.

Figure 4.3 - Prompt: a {green, red, blue} frog.

Conclusion. What’s Next?

In conclusion, with effective combined usage of commands, parameters, and features, you can greatly enhance the quality of your imagination and expand your creativity to levels you never thought you could before. Commands in Midjourney offer versatile tools for interacting with the bot, enabling users to create, blend, and describe images, as well as customize settings and access account information. Parameters play a crucial role in tailoring the image generation process to your creative vision, allowing for adjustments in aspect ratio, rendering quality, style, and more. Features allow for closer enhancements and edits in order to finalize generated images. You even learned some new ways of structuring prompts as well.

After delving into Midjourney for ten months through lively discussions on the Midjourney Discord, engaging with various Facebook groups dedicated to Midjourney, diving into prompt-building books, and devouring online articles, my understanding has grown significantly. Yet, I can't help but feel like I've only just begun to uncover the depths of this incredible AI. Every day brings fresh insights as I experiment with different techniques to customize Midjourney to suit my needs. I believe that your journey with Midjourney will be just as enriching, bringing you the same sense of delight and satisfaction that it has brought me. I trust you've found this article beneficial as you continue to explore and innovate on the Midjourney platform.