Midjourney Tutorial for Beginners

Basic Settings

Sending the /settings command to Midjourney Bot reveals configurable options:

Model Version Control

The first dropdown menu selects Midjourney’s image generation models, which differ in artistic tendencies and parameter compatibility:

• Midjourney Model: Native model for photorealistic and stylized images. Default version is V6.1.
• Niji Model: Collaborative model with SpellBrush, optimized for anime/illustration styles. Current version: V6.

Stylization Control

The second row controls artistic interpretation strength (higher values = more stylized results):

• RAW Mode: Equivalent to --style raw. Reduces AI stylization, available only in Midjourney Model V5.1/V5.2.
• Stylize low: Weak stylization, equals --style 50
• Stylize med: Med stylization, equals --style 100 (default)
• Stylize high: Strong stylization, equals --style 250
• Stylize very high: Extra strong stylization, equals --style 750

Personalization

The Personalization feature allows AI to learn creators’ preferences, reducing repetitive style explanations in prompts.
Prerequisite: Rate at least 200 images on Midjourney’s official website to train your preference profile.
Usage: Append --p to prompts after enabling this feature.

Privacy Settings

In Public mode, all generated images are publicly visible in Midjourney’s gallery. Only Pro Plan subscribers can disable this mode.

Remix Settings

• Remix mode: Allows editing prompts during Upscale/Variation/Reroll operations. Disabled by default.
• Strong Variation Mode: Maximizes output diversity (requires Remix mode).
• Subtle Variation Mode: Minimizes output changes (requires Remix mode).

GPU Modes

• Fast mode: Prioritizes speed (consumes Fast GPU time; ~seconds per image).
• Relax mode: Queued generation (no Fast time cost; ~seconds to minutes).
• Turbo mode: 4x faster than Fast mode but consumes 2x Fast time per task.

Command Reference

Basic Commands

• /settings: View/update configurations
• /help: Display help documentation
• /info: Show version details
• /about: Midjourney introduction
• /quit: Exit Midjourney

Image Generation Commands

• /imagine: Generate images
• /prompt: Set image description
• /seed: Specify random seed
• /variations: Create image variants
• /upscale: Enhance image resolution
• /style: Apply artistic style
• /quality: Adjust rendering quality
• /aspect: Set aspect ratio
• /mask: Use image masks

Advanced Commands

• /blend: Merge 2-5 images
• /inpaint: Image restoration
• /edit: Modify existing images
• /animate: Create animations
• /loop: Generate looping animations
• /slerp: Image interpolation
• /checkpoint: Save model state
• /restore: Load model state

Miscellaneous Commands

• /join: Enter chat room
• /leave: Exit chat room
• /mute: Silence notifications
• /unmute: Enable notifications
• /clear: Delete chat history

Usage Notes

• Commands are case-sensitive
• Commands can be combined
• Effects depend on prompt structure and parameter order

Prompt Practice

Basic Structure

A complete Midjourney prompt includes:

1. Command: Only /imagine accepts prompts
2. Prompt: Text descriptions/image URLs
3. Parameters: Suffixed with --parameter value

Example:

1

a cute cat; american anime style --niji 6 --ar 1:1 --s 150  

Generates four progressively refined images:

Key Parameters

Format: Add space after prompt + --parameter value

General Parameters

• --seed: Random seed (0-4294967295)
• --chaos: Variation intensity (0-100)
• --quality: Detail level (0.25-1)
• --aspect: Aspect ratio (default 1:1)
• --style: Named artistic styles
• --no: Exclude elements (e.g., --no faces)

Advanced Parameters

• --iw: Image weight (0-2)
• --fast: Prioritize speed over quality
• --repeat: Batch generation count
• --relax: Quality-focused generation
• --stylize: Style intensity (0-1)
• --mask: Apply image mask

Additional Parameters

• --promptw: Multi-weight text prompts
• --v: Model version selector
• --upscaler: Upsampling algorithm
• --prompt_weights: Text weighting
• --cfg_scale: Style adjustment scale

Upscale & Variation Workflow

After generating four initial images:

Button Functions:

• U: Upscale selected image
• V: Generate variants from selected image
• Reroll: Regenerate with original prompt

Post-Upscale Options:

• Upscale (Subtle/Creative): Enhance resolution with minor/major detail adjustments
• Vary (Subtle/Strong): Generate low/high-variance variants
• Vary (Region): Edit specific areas
• Zoom Out: Expand canvas (2x, 1.5x, or custom)
• Pan Controls: Adjust composition with arrow keys

For advanced techniques, see: 《Midjourney Advanced Guide》