Commands and parameters in Discord
Using Midjourney through Discord offers more than just generating images—it unlocks deeper control and flexibility through specific commands and parameters. In this section, we’ll go beyond the brief overview you encountered in Chapter 2 and dive into how commands and parameters work on Discord, how to write them correctly, and what each option means at a high level. This foundation will help you make the most of the advanced techniques introduced throughout Chapter 3.
What is a Discord command and parameter?
In Midjourney’s Discord interface, you interact with the model primarily through text prompts. But these prompts are more than just descriptive sentences—they can include special commands and parameters that modify how the model generates your image. A command usually begins with a slash (/), while a parameter starts with two dashes (--) and provides specific instructions about things like style, size, randomness, or aspect ratio.
Here’s a quick comparison:
- Command (e.g.,
/imagine): Triggers a specific Midjourney action. - Parameter (e.g.,
--ar 16:9): Refines or customizes the output of a command.
Commands are like “verbs” and parameters are like “adjectives” that shape your image generation request.
Discord command list
Let’s look at the primary commands used in Midjourney. In most cases, you’ll work with just one or two, but knowing what each does gives you a better grasp of what’s possible.
/imagine
This is the heart of the Midjourney experience. It generates new images based on a text prompt you provide. It also supports parameters like --v, --ar, and more to refine the output.
As we have already explained earlier, this command opens a field labeled prompt, where you type the description of the image you want. It’s the most used command in Midjourney and supports all major parameters.
/blend
This command lets you upload multiple images to combine them into a single, unified composition. It’s ideal for merging styles, themes, or subjects.
After selecting /blend, you’ll be prompted to upload 2–5 images. Midjourney blends them into a single output by interpreting shared patterns, colors, or themes. Blending works especially well for fashion, mood boards, or collage-style visuals.
We’ll explore the blend feature later in this chapter.
/describe
This is the reverse of /imagine. You upload an image, and Midjourney attempts to write a prompt that describes it. It’s helpful when you want to extract prompt ideas from a reference image. You can use this feature to generate prompt suggestions from your past outputs or found imagery, which you can then remix or build on.
By uploading an image…
Midjourney Bot returns the reverse engineered prompts from the image.
/settings
This command opens a clickable menu in Discord where you can adjust your default options.
These include:
- Model version (
v6,niji, etc.) - Style strength (
raw,default) - Quality level (
--q 1,--q 2) - Fast vs. Relax mode
Using /settings helps reduce the need to repeatedly add parameters like --v 6 or --style raw manually.
/ask
This command is less about image generation and more of a support feature. You can type a question, and Midjourney’s bot will try to assist—whether it’s about commands, features, or troubleshooting. For example, if you ask “What does the --seed parameter do?”
Midjourney Bot returns an answer.
Discord parameter list
Parameters are powerful tools that let you take control over how Midjourney interprets your prompts. They begin with -- and are typically added at the end of your /imagine command. Below, we’ve grouped the most relevant parameters into categories that match their function. You’ll explore each in more depth later in this chapter, but this overview will help you begin using them effectively right away.
Stylization and quality: Using --stylize, --raw, and --quality
These parameters define how polished, artistic, or faithful your images appear.
--stylizeor--s: Controls how strongly Midjourney applies its built-in artistic style. Higher values produce more decorative or abstract results.--raw: Bypasses Midjourney’s default stylistic influence, giving you a more literal interpretation of your prompt.--qualityor--q: Adjusts rendering time and detail.--q 2generates finer detail, while--q 1is standard.
Embracing randomness: Using --weird, --chaos, and --seed
These parameters influence how varied, unconventional, or repeatable your outputs are.
--chaosor--c: Increases variation in composition and interpretation. Higher values mean less predictability and more experimentation.--weirdor--w: Introduces strange or surreal interpretations—useful for creative exploration and avant-garde results.--seed: Locks in the random starting point for generation. Use it to reproduce the same image or create controlled variations.
Image-based prompting, blending, and style control
These parameters control how images, styles, or references influence the result.
--iw(Image Weight): Adjusts the influence of image prompts relative to text prompts. Higher values mean more visual influence.--sref(Style Reference): Matches the style of a previous image. Ideal for consistent aesthetics across different prompts.--oref(Omni Reference): Used to mimic the appearance or form of a person or object. Replaces older Character Reference features in Version 7.
Fine-tuning outputs: Using --no, --tile, and --repeat
These parameters help you guide composition or output structure more precisely.
--no: Excludes unwanted elements from the image (e.g.,--no textto remove text-like shapes).--tile: Creates a seamless, repeating texture—great for backgrounds and pattern design.--repeator--r: Repeats a prompt multiple times in one job, generating varied outputs without rewriting.
Power features: Profile settings and tuners in Discord
These options give you more control over personalization and visibility.
--profileor--p: Applies a saved style preset (known as a Tuner) to your prompt, helping maintain consistent visual identity.--stealth: Keeps generated images private on your Midjourney gallery. Available only to Pro users.--public: Publishes your image to your gallery—even if your default setting is private.
Additional controls: Version, speed, and testing tools
These modifiers manage which model version is used, how quickly images render, and how efficiently you test prompts.
--versionor--v: Selects the model version, such as--v 6or--nijifor anime-specific results.--fast,--relax, and--turbo: Toggle GPU usage modes: Fast for quick results, Relax for slower but unlimited rendering, and Turbo for priority speed (for eligible plans).
Mastering commands and parameters in Discord is the key to unlocking Midjourney’s full creative power. While the /imagine command gets you started, it’s the careful use of parameters like --ar, --stylize, --quality, and others that allow you to shape your vision precisely. Think of this setup as your creative toolkit—one that grows even more powerful as you layer these elements together.