Midjourney Commands & Prompt Reference
38 commands & parameters · 2 categories · Source: Midjourney Docs — Parameter List
Midjourney is driven entirely by typed input: slash commands start jobs in Discord, and parameters appended to the end of a prompt control aspect ratio, style strength, references, seeds, speed modes, and video options. This page combines the Discord command set with the current official parameter list.
Parameter placement matters: parameters go after the prompt text, with a space before the double dash and no punctuation inside them (vibrant poppies --ar 2:3). References are the modern workhorses: --sref matches a look, --sw controls how strongly, and --oref carries a subject's likeness in V7.
All Commands
Discord Slash Commands (7)
| Command / Syntax | What it does | Notes |
|---|---|---|
| /imagine | Generate image | Generate an image from a text prompt. |
| /describe | Describe image | Describe an uploaded image in text. |
| /blend | Blend images | Blend 2-5 images together. |
| /settings | Settings | Configure model version, quality, etc. |
| /info | Account info | Check subscription and usage. |
| U1~U4 | Upscale | Upscale the selected image. |
| V1~V4 | Create variation | Generate a variation of the selected image. |
Prompt Parameters (31)
| Command / Syntax | What it does | Notes |
|---|---|---|
| --ar <w:h> | Aspect Ratio: change the image shape from the default square | Also: --aspect. Example: --ar 2:3. Place after the prompt text. |
| --chaos <0-100> | Vary the results more between grid images | Also: --c. Higher values give more unexpected compositions. |
| --oref <url> | Omni Reference: use a person's likeness or an object's form | Replaces Character Reference in V7. |
| --no <object> | Tell Midjourney what you do not want in the image | Example: --no plants removes plants from results. |
| --profile | Personalization: apply your personalized profile or moodboard style | Also: --p. |
| --quality <value> | Control detail level and processing time | Also: --q. |
| --repeat <n> | Generate multiple image sets from a single prompt | Also: --r. Subscription tier limits apply. |
| --seed <number> | Fix the starting noise for testing and experimenting | Same seed and prompt reproduce similar results. |
| --stealth | Make your creations private on the Midjourney website | Availability depends on your plan. |
| --public | Make your creations public on the Midjourney website | Counterpart to --stealth. |
| --raw | Raw Mode: reduce automatic beautification for more literal results | Style-raw switch for the current model. |
| --stylize <value> | Control how strongly Midjourney's aesthetic is applied | Also: --s. Low values follow the prompt more literally. |
| --sref <url> | Style Reference: match the look and feel of another image | Combine with --sw to control strength. |
| --sw <value> | Style Weight: control the strength of Style References | Used together with --sref. |
| --sv <value> | Select the style reference algorithm version | Used together with --sref. |
| --tile | Create seamless repeating patterns | Output tiles edge-to-edge for textures. |
| --version <n> | Switch between Midjourney model versions | Also: --v. Example: --v 7. |
| --niji | Use the model focused on anime and Eastern aesthetics | Niji model family switch. |
| --draft | Generate draft images in V7 at half the GPU cost | Faster, lower-cost exploration. |
| --weird <value> | Make images quirky and unconventional | Also: --w. |
| --iw <value> | Image Weight: control the impact of image prompts | Higher values follow the linked image more closely. |
| --fast | Switch GPU speed to Fast Mode for this job | Consumes fast GPU time. |
| --relax | Switch GPU speed to Relax Mode for this job | No fast-hour cost; queued generation. |
| --turbo | Switch GPU speed to Turbo Mode for this job | Fastest option; higher GPU cost. |
| --hd | Generate V8.1 images at higher resolution (2048px) | Counterpart: --sd for standard 1024px. |
| --sd | Generate V8.1 images at standard resolution (1024px) | Counterpart: --hd for 2048px. |
| --video | Generate videos in Discord | Video generation switch for Discord workflows. |
| --motion low | Low motion setting for video generations | Counterpart: --motion high. |
| --motion high | High motion setting for video generations | More movement between video frames. |
| --loop | Create looping video generations | Combine with --end to set a custom end frame. |
| --bs <n> | Batch Size: number of videos generated per video prompt | Controls video batch output count. |
Recipes
Our own notes on combining the commands above. Every token used here appears in the reference tables on this page.
Reproduce a composition you liked at a different aspect ratio
--seed <number> --ar 16:9
A fixed seed plus the same prompt reproduces a similar starting point, so changing only the aspect ratio isolates that one variable. Without the seed, changing anything reshuffles the whole image.
Carry one image's look across a whole set
--sref <url> --sw <value>
--sref matches the look and feel of a reference image; --sw controls how strongly it applies. Start at a moderate weight and raise it only if the style is not carrying - a maximum weight tends to override the prompt's subject.
Keep the same person or object across images in V7
--oref <url>
Omni Reference carries a person's likeness or an object's form, and replaces the older Character Reference workflow. Combine it with --sref when you need both a consistent subject and a consistent style.
Explore cheaply, then commit resolution to the winner
--draft → pick → --hd
--draft generates at half the GPU cost in V7, which is the right mode for the first twenty ideas. Once one works, regenerate it with --hd for 2048px output. Spending full cost on exploration is the most common way to burn fast hours.
Make the result follow your words rather than the house style
--stylize <low value> --raw
--stylize controls how strongly Midjourney's own aesthetic is applied; lower values follow the prompt more literally. --raw reduces automatic beautification further. Pair with --no <object> to remove something that keeps appearing.
Generate several variations of one prompt in a single job
--repeat <n> --chaos <0-100>
--repeat runs the prompt multiple times; --chaos widens how much the results differ from each other. Low chaos with repeat gives you near-misses of the same idea, high chaos gives genuinely different takes.
Frequently Asked Questions
Where do parameters go in a Midjourney prompt?
Always at the end, after the prompt text, with a space before the dashes and no punctuation inside the parameter (vibrant poppies --ar 2:3).
What replaced Character Reference in V7?
Omni Reference: --oref takes a reference for a person's likeness or an object's form, replacing the older Character Reference workflow.
How do I make results more or less literal?
--stylize controls how strongly Midjourney's aesthetic is applied (lower is more literal), and --raw reduces automatic beautification further.
Every entry on this page is a literal, typeable token taken from the official documentation linked above. Command sets in AI tools change quickly - the in-product command menu is always authoritative for your installed version.