How to Use Midjourney AI: A Step by Step Tutorial

Getting Started

Step 1: Join Discord

Midjourney operates through Discord. Go to midjourney.com and click "Join the Beta" to access the Discord server.

Step 2: Subscribe

Use the /subscribe command in any Discord channel. Choose a plan:

Plan	Monthly	Images	Commercial Use
Basic	$10	~200	Yes
Standard	$30	Unlimited (slow) / 15h fast	Yes
Pro	$60	Unlimited (slow) / 30h fast	Yes
Mega	$120	60h fast	Yes

Step 3: Generate Your First Image

Go to a #newbies-* channel or your private server. Type:

/imagine prompt: a serene mountain lake at sunset, photorealistic, 8k, cinematic lighting

Midjourney will process and return four variations.

The Prompt Structure

[Subject] + [Details] + [Style] + [Technical Parameters]

Example:
a futuristic city with neon lights, rain on streets, cyberpunk style,
highly detailed, cinematic lighting --ar 16:9 --v 6

Elements of a Good Prompt

• Subject: What is the main focus? (a cat, a landscape, a product)
• Medium: Photo, painting, 3D render, sketch
• Environment: Setting/background
• Lighting: Golden hour, cinematic, studio, neon
• Color palette: Pastel, monochrome, vibrant
• Mood: Serene, dramatic, mysterious
• Composition: Close-up, wide angle, bird's eye view

Parameters Reference

Parameter	Example	Effect
--ar	--ar 16:9	Aspect ratio (width:height)
--v	--v 6	Version (6 is latest)
--s	--s 250	Stylize (0-1000, higher = more artistic)
--c	--c 20	Chaos (0-100, higher = more diverse)
--iw	--iw 2	Image weight when using reference images
--no	--no text,watermark	Negative prompt — what to exclude
--seed	--seed 12345	Reproduce similar results
--stylize	--stylize 500	Alternative to --s

Advanced Techniques

Image Referencing

Upload an image to Discord, copy its URL, and use it in your prompt:

/imagine prompt: [image URL] a cat sitting on a chair --iw 1.5

Remix Mode

After generating, use Vary (Region) to modify specific areas, or use Remix to change the prompt while keeping the composition.

Blending

/blend

Select two images and Midjourney blends them.

Pro Tips

1. Be specific — "a dog" vs "a golden retriever puppy sleeping on a blue velvet cushion"
2. Use style references — "in the style of Studio Ghibli" or "in the style of Annie Leibovitz"
3. Iterate — Upscale the best result, then vary it
4. Keep prompts concise — 10-15 words is ideal
5. Use parameters wisely — Higher stylize for art, lower for realism

Common Mistakes

• Too vague — "a nice picture of something beautiful" gives poor results
• Overloading — Too many subjects confuse the model
• Wrong aspect ratio — Default is square (1:1); specify --ar for landscapes/portraits
• Ignoring version --v 5.2 and --v 6 produce very different results