Midjourney Cheat Sheet

Reviewed by Marouen Arfaoui · Last tested April 2026 · 157 tools tested

Last updated: April 2026

Quick Facts

Pricing

Paid subscription only, starting at $10/month.

Free Plan

No. A limited trial (~25 images) is sometimes offered to new users.

Rating

4.8/5

Best For

Artists and creators who prioritize artistic style, composition, and a unique aesthetic over strict photorealism or control.

Key Features

✓
Discord-Based Interface
I work entirely within Discord. This creates a unique, social environment where I can see others' prompts and results, which is fantastic for learning.
✓
Stylized Aesthetic
Midjourney's core strength is its beautiful, painterly, and often cinematic default style. In my experience, it consistently produces more 'artistic' images than competitors.
✓
Prompt Understanding (--style raw)
Using --style raw gives me more literal interpretation of my prompts. It's less opinionated, which I prefer for technical or specific concepts where I need control.
✓
Upscaling & Variations
After generating a grid, I can upscale a chosen image to high resolution or create four new variations of it. This iterative refinement is my core workflow.
✓
Aspect Ratio Control (--ar)
I use --ar 16:9 for landscapes, --ar 9:16 for phone wallpapers, and --ar 1:1 for social posts. It's crucial for framing the final image correctly from the start.
✓
Stylize Parameter (--s)
The --s parameter (0-1000) controls how strongly Midjourney applies its artistic flair. Lower values follow my prompt closely; higher values are more creative and surprising.
✓
Chaos Parameter (--c)
When I'm feeling stuck, I add --c 100 to my prompt. This introduces more randomness into the initial grid, often leading to unexpected and inspiring compositions.
✓
Weirdness Parameter (--weird)
For truly surreal and abstract art, --weird is my go-to. It pushes the AI to break conventional patterns, resulting in bizarre and fascinating imagery.
✓
Remix Mode
With Remix enabled, I can change the prompt when making variations. This lets me guide an image in a new direction, like adding a hat to a character mid-process.
✓
Pan & Zoom (Vary Region)
I can expand an image beyond its borders or zoom into a specific area and regenerate it. This is essential for creating wide-format scenes or refining details.
✓
Image Prompts
I can upload a reference image alongside my text prompt. Midjourney blends the style and content, which is perfect for creating consistent character portraits or art styles.
✓
Describe Command
If I see an image I like but don't know the prompt, I can use /describe. It gives me four text interpretations, which is a brilliant way to reverse-engineer techniques.

Tips & Tricks

TIP

Start simple. A prompt like 'a wise old wizard, detailed portrait, studio lighting' works better than an overly complex paragraph.

TIP

Use commas to separate concepts, not periods. Think of it as giving the AI a list of keywords to combine, not writing a sentence.

TIP

Learn key parameters like --ar for aspect ratio and --s for stylization. They are more powerful than adding more descriptive words.

TIP

Study the community feed in Discord. Seeing others' prompts and results is the fastest way to learn advanced prompting techniques.

TIP

Don't settle on the first grid. Use the 'V' buttons to create variations of your favorite image to explore the concept further.

TIP

For consistent characters, generate an image you like, then use its seed number (--seed) in future prompts to maintain likeness.

TIP

Add '--no' to exclude elements. For example, 'a clean kitchen --no people, clutter, mess' helps steer the AI away from unwanted items.

Common Commands

/imagine prompt: [your text here]

The primary command to generate a new grid of 4 images based on your text description.

U1, U2, U3, U4

Upscales the selected image (1-4) from the initial grid to a higher resolution.

V1, V2, V3, V4

Creates a new grid of 4 variations based on the selected image (1-4) from the initial grid.

/settings

Opens a menu to toggle key modes like Remix, Fast/Relaxed generation, and the default model version.

/describe

Upload an image, and Midjourney will suggest four possible text prompts that could have created it.

Limitations

-It struggles terribly with rendering accurate text and specific logos; expect gibberish.
-Hands, fingers, and complex anatomy are often flawed, requiring multiple generations or inpainting to fix.
-The Discord interface can feel chaotic and is not ideal for organizing a large library of personal work.
-You have no direct, private web interface; all basic plan images are public in the community gallery.
-It has a distinct 'Midjourney look' that can be hard to break away from without advanced prompting.

Alternatives

DALL-E 3 (via ChatGPT or Bing)Stable Diffusion (via UI like ComfyUI)Adobe Firefly

→

Midjourney TutorialFull step-by-step guide

→

Frequently Asked Questions

Is there a free trial for Midjourney?+

Officially, no permanent free plan exists. However, Midjourney occasionally offers a limited trial of about 25 image generations to new users who join their Discord server. This is not always available. For consistent use, you will need a paid subscription starting at $10 per month.

Can I use Midjourney images commercially?+

Yes, if you are a paying subscriber on a Standard, Pro, or Mega plan, you own the assets you create and can use them for commercial purposes, including selling them. You must review the full Terms of Service, but broadly, you have broad commercial rights. The Basic plan has some limitations on high-volume commercial use.

How do I get a square (1:1) image?+

The default aspect ratio is 1:1, so you don't need to do anything for a square. If you've changed settings, simply add '--ar 1:1' to the end of your prompt. For other formats, use '--ar 16:9' for widescreen or '--ar 2:3' for a portrait orientation.

What does 'Fast Mode' vs. 'Relaxed Mode' mean?+

Fast Mode uses your monthly GPU time allotment and generates images in seconds. Relaxed Mode is free/unlimited for Standard+ plans but queues your job, which can take minutes during peak times. I use Fast for active brainstorming and Relaxed for overnight batches.

How can I make a consistent character across multiple images?+

The most reliable method is to generate a good base image, copy its seed number (found in the job's info or by reacting with an envelope), and use '--seed [number]' in future prompts along with a similar description. Using image prompts of the character also helps maintain consistency.

Was this helpful?