How to Use Midjourney: Complete Beginner’s Guide (2026)

Midjourney has become one of the most talked-about AI image generators on the planet. From hobbyists making desktop wallpapers to professional designers prototyping brand visuals, millions of people rely on Midjourney to turn text descriptions into striking images within seconds.

But if you have never used Midjourney before, the learning curve can feel steep. Where do you even start? What does /imagine do? What are all those parameters like --ar and --v?

This guide walks you through everything from creating your first image to fine-tuning advanced parameters. By the end, you will be generating images that look like they came from a professional studio.

What Is Midjourney?

Midjourney is an AI-powered image generation tool developed by an independent research lab based in San Francisco. You give it a text prompt describing what you want to see, and within seconds it produces four image variations based on your description.

Unlike some competitors, Midjourney is known for a distinctive aesthetic quality that leans toward painterly, cinematic visuals. That said, it handles photorealism, illustration, and abstract styles equally well when you know how to prompt it correctly.

What You Can Create with Midjourney

• Concept art for games, films, and books
• Marketing visuals for social media, ads, and presentations
• Product mockups and packaging designs
• Architectural visualizations and interior design concepts
• Personal art including portraits, landscapes, and fantasy scenes
• Logo ideas and brand identity exploration
• AI video clips (new in 2025)

Midjourney Pricing Plans (2026)

Before you get started, you need a paid subscription. Midjourney suspended its free trial in April 2023, so there is no free tier available as of 2026.

Plan	Monthly Price	Annual Price	Fast GPU Hours	Relax Mode	Stealth Mode	Concurrent Jobs
Basic	$10/mo	$8/mo	~3.3 hrs	No	No	3
Standard	$30/mo	$24/mo	~15 hrs	Yes	No	3
Pro	$60/mo	$48/mo	~30 hrs	Yes	Yes	12
Mega	$120/mo	$96/mo	~60 hrs	Yes	Yes	12

Which plan should you pick? If you are just starting out, the Basic plan at $10 per month gives you enough room to experiment. Once you find yourself running out of GPU hours regularly, the Standard plan is the sweet spot for most users because it unlocks unlimited Relax Mode generations.

Step 1: Create a Midjourney Account

Getting started with Midjourney takes just a few minutes.

Go to Midjourney.com and click “Sign Up” or “Join the Beta.”

Create your account using your email address or sign in with Discord or Google.

Choose a subscription plan from the pricing page. You will need to enter payment information.

Access the Midjourney web app at midjourney.com or use the Discord bot in the official Midjourney Discord server.

!Screenshot placeholder: Midjourney sign-up page with subscription options

Midjourney originally operated exclusively through Discord. While the Discord bot is still fully functional, most new users now prefer the web interface at midjourney.com, which offers a cleaner experience with built-in image browsing, editing tools, and prompt history.

Step 2: Generate Your First Image with /imagine

The core command in Midjourney is /imagine. This is how you tell the AI what to create.

Using the Web App

Log in to midjourney.com

Click the “Create” tab in the navigation bar

Type your prompt in the text box at the bottom of the screen

Press Enter or click the arrow button to submit

Using Discord

Join the Midjourney Discord server or invite the Midjourney bot to your own server

Go to any bot channel (like #newbies)

Type /imagine followed by your prompt

Press Enter and wait for the bot to generate your images

!Screenshot placeholder: The /imagine command being used in the Midjourney web app

Your First Prompt

Start simple. Type something like:

/imagine a golden retriever sitting in a field of wildflowers at sunset

Within 30 to 60 seconds (depending on your plan and current server load), Midjourney will generate a grid of four image variations. From there, you can:

• U1, U2, U3, U4 — Upscale a specific image to full resolution
• V1, V2, V3, V4 — Create new variations based on a specific image
• Re-roll — Generate four completely new variations from the same prompt

Step 3: Write Better Prompts

The quality of your output depends heavily on how you write your prompt. Midjourney does not require overly technical language, but being specific about what you want makes a significant difference.

Anatomy of a Good Prompt

A strong Midjourney prompt usually includes several of these elements:

Subject — What is the main focus? (a knight, a cat, a futuristic city)

Setting/Environment — Where is it? (in a forest, on a rooftop, underwater)

Lighting — What is the mood? (golden hour, neon lights, overcast)

Style — What aesthetic are you going for? (oil painting, photography, anime)

Camera/Perspective — What angle? (close-up, aerial view, wide shot)

Details — Specific elements that matter (wearing a red scarf, holding a lantern)

Prompt Examples: From Basic to Advanced

Basic:

a cat

Better:

a tabby cat sitting on a windowsill, soft morning light, cozy apartment

Advanced:

a tabby cat sitting on a windowsill overlooking a rainy Tokyo street, soft morning light filtering through sheer curtains, shot on 35mm film, warm color palette, nostalgic mood

Notice how each version adds more context. You do not need to write an essay, but giving Midjourney two or three descriptive layers dramatically improves results.

Words That Influence Style

Certain keywords steer Midjourney toward specific looks:

Keyword	Effect
photorealistic	Realistic photograph look
cinematic	Film-like composition and lighting
oil painting	Traditional painted aesthetic
watercolor	Soft, translucent painted look
anime	Japanese animation style
isometric	3D technical illustration angle
minimalist	Clean, simple design
baroque	Ornate, dramatic classical style
cyberpunk	Neon-lit, futuristic dystopia
studio lighting	Clean, professional portrait look

Step 4: Master Midjourney Parameters

Parameters are modifiers you add to the end of your prompt to control technical aspects of the generation. They always start with two dashes (--) and go after your description text.

–ar (Aspect Ratio)

By default, Midjourney creates square images (1:1). The --ar parameter lets you change the width-to-height ratio.

/imagine a panoramic mountain landscape at dawn --ar 16:9

Common aspect ratios:

Ratio	Use Case
1:1	Instagram posts, profile pictures
4:3	Presentations, classic photo prints
3:2	Standard photography
16:9	YouTube thumbnails, desktop wallpapers
9:16	Instagram Stories, TikTok, phone wallpapers
2:3	Pinterest pins, portrait photos
21:9	Ultra-wide cinematic banners

!Screenshot placeholder: Side-by-side comparison of images with different aspect ratios

–v (Version)

Midjourney has released several model versions, each with its own strengths. The current default is V7, released in April 2025.

/imagine a medieval castle on a cliff --v 7

Version	Strengths
V7 (default)	Best coherence, hands, text rendering, Draft Mode
V6.1	Strong detail, 25% faster than V6
V6	Good photorealism, reliable
Niji 7	Anime and illustration specialist

Unless you have a specific reason to use an older version, stick with V7. It handles prompts with noticeably better precision than previous versions, especially for human anatomy, text within images, and complex compositions.

–style raw

By default, Midjourney adds its own creative interpretation to your prompts. The --style raw parameter turns off this automatic styling, giving you more literal results.

/imagine a plain white coffee mug on a wooden table --style raw

This is especially useful when you want photorealistic images or need precise control over the look. With simple prompts, Raw Mode tends to produce more realistic, photo-like output.

–stylize or –s (Stylize)

The --stylize parameter (shorthand --s) controls how much artistic flair Midjourney adds. It ranges from 0 to 1000, with a default of 100.

/imagine a forest path in autumn --s 250

• Low values (0-50): Closely match your prompt description with minimal artistic interpretation
• Default (100): Balanced mix of accuracy and style
• High values (250-750): More artistic and evocative, but may drift from your exact description
• Very high (750-1000): Heavily stylized, sometimes abstract

–chaos or –c

The --chaos parameter (0 to 100) controls how different the four variations in your grid will be from each other.

/imagine abstract art inspired by ocean waves --c 50

• Low chaos (0-20): All four images look similar
• Medium chaos (30-60): Noticeable variety between images
• High chaos (70-100): Wildly different interpretations

This is great for brainstorming when you want to explore different directions from a single prompt.

–no (Negative Prompting)

The --no parameter tells Midjourney what to leave out of the image.

/imagine a peaceful village street at night --no cars people

This removes cars and people from the result. It does not guarantee perfect removal, but it significantly reduces the chances of those elements appearing.

–seed

Every image Midjourney generates uses a random seed number. If you find an image you like and want to create similar variations, you can specify the same seed.

/imagine a cyberpunk alley in the rain --seed 12345

Using the same seed with the same prompt produces very similar (though not identical) results across generations.

–q (Quality)

The --q parameter adjusts how much processing time Midjourney spends on your image.

/imagine a detailed dragon sculpture --q 2

Higher quality values use more GPU time but produce more detailed results.

Step 5: Use /blend to Combine Images

The /blend command lets you upload two to five images and merge them together. Midjourney analyzes the visual elements of each image and creates a new composition that blends their characteristics.

How to Use /blend

Type /blend in the prompt box

Upload 2 to 5 images when prompted

Optionally add parameters like --ar

Submit and wait for the result

!Screenshot placeholder: The /blend command interface showing two uploaded images

Use cases for /blend:

• Combining a photo of a person with an artistic style reference
• Merging architectural styles from different buildings
• Creating hybrid animal or creature designs
• Blending color palettes from different images

Tips for better blends:

• Use images with similar compositions for more coherent results
• The blend works best with 2-3 images; more than that can get muddy
• Try blending a photo with an art style reference for striking results

Step 6: Edit and Refine Your Images

Once you have generated an image you like, Midjourney offers several ways to refine it.

Upscaling

Click the U button (U1 through U4) under your image grid to upscale a specific image. This produces a higher-resolution version suitable for printing or professional use.

Variations

Click the V button (V1 through V4) to create new images based on the composition and style of a specific result. This is useful when you like the direction but want to explore slight differences.

Panning and Zooming

After upscaling an image, you can:

• Pan in any direction (left, right, up, down) to extend the image canvas
• Zoom out to reveal more of the scene around your original image

These features are powerful for creating wider compositions from an initial generation.

Inpainting (Vary Region)

The Vary Region tool lets you select a specific area of an upscaled image and regenerate just that portion while keeping the rest intact.

Upscale an image

Click “Vary (Region)”

Use the selection tool to highlight the area you want to change

Write a new prompt describing what should appear in that region

Submit and review the results

This is incredibly useful for fixing small issues like awkward hands, adding details, or changing specific elements without regenerating the entire image.

Pro Tips for Getting the Most Out of Midjourney

After generating thousands of images, experienced users tend to rely on a handful of strategies that consistently produce better results.

1. Reference Real Artists and Photographers

Including the name of a well-known artist or photographer in your prompt steers the style effectively. For example, adding “in the style of Ansel Adams” pushes the output toward dramatic black-and-white landscape photography.

2. Specify Camera and Lens Details

For photorealistic results, include camera terminology:

/imagine portrait of an elderly fisherman, shot on Canon EOS R5, 85mm lens, f/1.4, shallow depth of field, natural light

3. Use Aspect Ratio Strategically

Do not default to square images. Think about where the image will be used and match the aspect ratio accordingly. A YouTube thumbnail needs --ar 16:9, while a phone wallpaper needs --ar 9:16.

4. Iterate with Variations

Rarely does the first generation nail exactly what you want. Use V buttons to create variations, then refine further. The best Midjourney images often come from 3 to 5 rounds of iteration.

5. Save Your Best Seeds

When you generate something you love, note the seed number. You can find it by reacting to your image with an envelope emoji in Discord, or checking the image details in the web app. This lets you recreate similar compositions later.

6. Combine Multiple Parameters

Parameters stack. A prompt like this uses several at once:

/imagine a bioluminescent underwater cave, mysterious atmosphere, volumetric lighting --ar 16:9 --s 300 --c 30 --v 7

7. Use Negative Prompts to Clean Up

If your images keep including unwanted elements (text, watermarks, blurry areas), add --no text watermark blurry to your prompt.

8. Explore Relax Mode

If you are on the Standard plan or above, switch to Relax Mode for unlimited generations at a slower speed. This is perfect for experimentation when you do not need results instantly.

Midjourney V7: What Changed

Midjourney V7, which became the default model in June 2025, brought several meaningful improvements over V6:

• Better hands and anatomy — The most common complaint about AI art has been malformed hands. V7 handles them significantly better.
• Improved text rendering — You can now include text in images with reasonable accuracy. Not perfect, but a major step forward.
• Draft Mode — A new fast-generation option for quick previews before committing GPU time to a full render.
• Omni Reference — A new referencing system that lets you guide the AI with more precise visual references.
• Richer textures — Surfaces, fabrics, and materials look more detailed and realistic.

AI Video Generation

In 2025, Midjourney expanded into AI video generation. You can now create short video clips (roughly 10 to 15 seconds) from text prompts or from existing still images.

A few things to keep in mind:

• Video generation consumes significantly more GPU time than still images, often around eight times the cost per generation
• Quality varies, and complex motion scenes are still challenging
• This feature works best for ambient scenes, slow camera movements, and simple animations
• For heavy video work, you may want the Pro or Mega plan

Frequently Asked Questions

Is Midjourney free to use in 2026?

No. Midjourney suspended its free trial in April 2023. As of 2026, all users need a paid subscription starting at $10 per month for the Basic plan.

Can I use Midjourney images commercially?

Yes. All paid plans include commercial usage rights, meaning you can use generated images in marketing materials, products, publications, and more. However, companies with over $1 million in annual gross revenue are required to subscribe to the Pro or Mega plan.

What is the difference between Fast Mode and Relax Mode?

Fast Mode uses your allocated GPU hours to generate images quickly (usually under a minute). Relax Mode, available on Standard plans and above, lets you generate unlimited images at a slower pace. Relax Mode images might take 1 to 10 minutes depending on server demand.

Do I need Discord to use Midjourney?

Not anymore. While Midjourney started as a Discord-only tool, the web app at midjourney.com now provides a full-featured interface for generating, browsing, and editing images. You can still use the Discord bot if you prefer that workflow.

How do I get consistent characters across multiple images?

This remains one of the trickier aspects of AI image generation. Your best options are using the same seed number, providing a reference image, or using the Omni Reference feature in V7. Describe the character in detail each time and keep core descriptors consistent across prompts.

Wrapping Up

Midjourney is one of those tools where the basics take five minutes to learn, but mastery can take months. The key is experimentation. Try different prompts, adjust parameters, explore styles you would never normally consider, and pay attention to which keywords produce the results you like most.

Start with the Basic plan to get your feet wet. Once you find yourself hitting generation limits and wanting more control, upgrade to Standard for unlimited Relax Mode. And if privacy matters for your work, the Pro plan with Stealth Mode is worth the investment.

The best way to learn Midjourney is to use it. Open up the web app, type a prompt, and see what happens.

You may also want to explore our roundup of Notion AI vs ChatGPT.

You may also want to explore our roundup of Perplexity vs Google comparison.

You may also want to explore our roundup of Notion AI alternatives.