Midjourney Prompting

Expert guidance for crafting effective Midjourney prompts using professional techniques and frameworks.

When to Use This Skill

Use this skill when users need:

• Converting natural language descriptions into effective Midjourney prompts
• Optimizing existing prompts for better results
• Guidance on Midjourney parameters (--ar, --s, --v, etc.)
• Understanding prompt structure and best practices
• Photography terminology and techniques for photorealistic images
• Style references and advanced prompting techniques

The 7-Element Framework

The most systematic approach to crafting Midjourney prompts. Use this framework to ensure comprehensive, effective prompts:

Element	Purpose	Examples
Subject	What's in the image	portrait, landscape, product, character, scene
Medium	Visual style/art form	photography, oil painting, digital art, watercolor
Environment	Setting/location	studio, forest, city street, underwater, space
Lighting	Light quality/direction	golden hour, studio lighting, dramatic shadows, soft diffused
Color	Color palette/mood	vibrant, muted, monochrome, warm tones, cool blues
Mood	Emotional tone	serene, energetic, mysterious, joyful, melancholic
Composition	Framing/perspective	close-up, wide angle, bird's eye view, rule of thirds

Quick structure:

[Subject] [Description], [Medium], [Environment], [Lighting], [Color palette], [Mood], [Composition] [Parameters]

Prompt Engineering Process

Step 1: Understand the User's Intent

Ask clarifying questions if needed:

• What's the main subject?
• What style are they aiming for? (photorealistic, illustrated, artistic)
• What mood or atmosphere do they want?
• What's the intended use? (social media, print, concept art)
• Any technical requirements? (aspect ratio, quality level)

Step 2: Transform Natural Language to Structured Prompt

Example transformations:

User request: "I need a professional headshot" Your prompt:

professional headshot, studio photography, neutral gray background, soft directional lighting, sharp focus, natural skin tones, confident expression, 85mm lens, shallow depth of field --ar 2:3 --style raw --v 7

User request: "Create a cyberpunk city scene" Your prompt:

cyberpunk city at night, neon lights reflecting on wet streets, flying cars, towering skyscrapers, heavy rain, purple and cyan color palette, cinematic atmosphere, wide angle shot --ar 16:9 --v 7

User request: "A cozy coffee shop interior" Your prompt:

cozy coffee shop interior, rustic wooden furniture, hanging plants, warm ambient lighting, customers reading and working, steaming cups, natural light from large windows, inviting atmosphere, architectural photography --ar 16:9 --v 7

User request: "Fantasy character design" Your prompt:

elven warrior character design, detailed armor with nature motifs, flowing cape, mystical forest background, ethereal lighting, emerald and gold color scheme, heroic pose, concept art style --ar 2:3 --v 7

Step 3: Add Appropriate Parameters

Core parameters to consider:

• --v 7 - Always use V7 for best quality (current version)

• --ar [ratio] - Aspect ratio based on use case:

  • --ar 16:9 - Landscape, presentations, YouTube thumbnails
  • --ar 1:1 - Social media posts, profile pictures
  • --ar 9:16 - Mobile, stories, vertical content
  • --ar 2:3 or 3:2 - Portraits, print photography
  • --ar 4:5 - Instagram posts

• --style raw - For photorealism (removes artistic interpretation)

• --s [0-1000] - Stylization level (default 100):

  • --s 0-50 - Very literal, photorealistic
  • --s 100-250 - Balanced (recommended for most)
  • --s 500-1000 - Very artistic, abstract

• --q [0.25-2] - Quality/detail level:

  • --q 1 - Standard (recommended, good balance)
  • --q 2 - Higher detail (slower, more expensive)

• --chaos [0-100] - Variation amount:

  • --chaos 0-20 - Consistent, predictable
  • --chaos 50-100 - Diverse, experimental

Advanced parameters:

• --seed [number] - For reproducible results
• --no [element] - Exclude unwanted elements (e.g., --no people --no text)
• --tile - For seamless patterns
• --sref [url] - Style reference image
• --cref [url] - Character reference image

See references/midjourney_parameters.md for complete parameter reference.

V7 Best Practices

DO:

• ✓ Be specific and descriptive about what you want to see
• ✓ Front-load important elements (put key concepts first)
• ✓ Use concrete visual terms, not abstract concepts
• ✓ Include lighting, environment, and atmosphere details
• ✓ Use --style raw for photorealism
• ✓ Trust V7's default quality settings

DON'T:

• ✗ Use junk words: "4K", "8K", "HD", "UHD", "high quality", "professional", "award-winning", "trending on artstation", "highly detailed", "octane render", "unreal engine"
• ✗ Over-explain or use overly long prompts (V7 is smart, be concise)
• ✗ Use negative prompts excessively (V7 understands intent better)
• ✗ Stack multiple similar descriptors ("beautiful gorgeous stunning")

V7 is smart enough - It understands context and intent. Focus on describing what you want to see, not quality metrics.

Photography Terminology

For photorealistic images, use precise photography terms:

Focal lengths:

• 14-24mm - Ultra wide, dramatic perspective, architecture
• 35mm - Wide, environmental, street photography
• 50mm - Standard, natural perspective, everyday
• 85mm - Portrait, flattering compression, headshots
• 135mm - Tight portrait, fashion, editorial
• 200mm+ - Telephoto, compression, wildlife, sports

Lighting:

• Studio lighting, softbox, beauty lighting
• Golden hour, blue hour, harsh midday sun
• Dramatic shadows, rim lighting, backlighting
• Soft diffused, overcast, natural window light
• Cinematic lighting, volumetric lighting

Camera bodies & film stocks:

• Hasselblad (medium format, high-end commercial)
• Leica (premium, street photography aesthetic)
• Canon 5D/1DX (professional photography)
• Kodak Portra 400 (warm, natural skin tones)
• Fuji Velvia (saturated, vibrant landscapes)
• Cinestill 800T (tungsten, cinematic)

See references/photography.md for comprehensive photography terminology.

Common Prompting Patterns

Pattern 1: Photorealistic Portrait

[Subject] portrait, [age/ethnicity/features], [expression], [clothing style],
[lighting], [background], [lens], shallow depth of field,
natural skin tones --ar 2:3 --style raw --v 7

Pattern 2: Architectural/Interior

[Space type], [style/period], [key features], [lighting conditions],
[atmosphere], [perspective], architectural photography --ar 16:9 --v 7

Pattern 3: Product Photography

[Product] on [surface], [setting/environment], [lighting setup],
clean composition, professional product photography,
[key features], --ar 4:5 --style raw --v 7

Pattern 4: Landscape/Environment

[Location/scene], [time of day], [weather conditions], [key features],
[color palette], [mood], [composition], landscape photography --ar 16:9 --v 7

Pattern 5: Character/Illustration

[Character description], [pose/action], [setting], [art style],
[color palette], [mood], [composition], [medium] --ar 2:3 --v 7

Advanced Techniques

Multi-Prompting with Weights

Use :: to separate concepts and assign weights:

sunset::2 ocean::1 sailboat::1 --v 7

This emphasizes "sunset" twice as much as ocean or sailboat.

Style References (--sref)

Apply the style of a reference image:

portrait of a woman --sref https://example.com/reference-style.jpg --sw 100 --v 7

• --sw [0-1000] controls style weight (default 100)

Character References (--cref)

Maintain consistent character appearance:

[character action/pose] --cref https://example.com/character.jpg --cw 100 --v 7

• --cw [0-100] controls character weight (default 100)

Negative Prompting

Exclude unwanted elements:

bedroom interior --no clutter --no windows --no people --v 7

Seed Consistency

For variations of the same concept:

1. Note the seed from a successful generation
2. Use --seed [value] in subsequent prompts
3. Modify other aspects while maintaining base consistency

Iterative Refinement

When users want to improve existing results:

1. Too dark/light? → Adjust lighting keywords:

   • Add: "bright", "well-lit", "sunny", "vibrant"
   • Or: "moody", "dim", "dramatic shadows"

2. Wrong style? → Clarify the medium/style:

   • Add specific art style: "oil painting", "watercolor", "digital art"
   • Or photography style: "35mm film", "Hasselblad", "cinematic"

3. Composition issues? → Add framing details:

   • "close-up", "wide shot", "bird's eye view"
   • "centered composition", "rule of thirds"

4. Want consistency? → Use the seed:

   • Get seed from successful image
   • Add --seed [value] to maintain base appearance

5. Need more/less stylization? → Adjust parameters:

   • Too artistic: lower --s value or add --style raw
   • Too literal: increase --s value

Quick Reference

Essential template:

[What] [doing what], [where], [lighting], [mood], [style] --ar [ratio] --v 7

For photos:

[Subject], [setting], [lighting], [lens], [aesthetic] --ar [ratio] --style raw --v 7

For art:

[Subject], [action/pose], [environment], [art style], [color palette], [mood] --ar [ratio] --v 7

Reference Documentation

This skill provides comprehensive reference materials:

1. prompt_engineering.md - Complete 7-Element Framework guide with examples
2. midjourney_parameters.md - All Midjourney parameters explained
3. photography.md - Photography terminology for photorealistic prompts

Best Practices Summary

1. Start with clarity - Understand what the user wants to see
2. Use the framework - Apply 7-Element Framework for systematic prompts
3. Be specific - Use concrete visual terms, not abstract concepts
4. Front-load importance - Key elements first
5. Choose right parameters - Match aspect ratio and style to use case
6. Trust V7 - Avoid quality junk words, V7 is already high quality
7. Iterate smartly - Refine based on results, use seeds for consistency
8. Photography terms for photos - Use precise technical terminology
9. Keep it focused - Concise, clear prompts work better than long ones
10. Test and refine - Prompting is iterative, adjust based on results

Example Workflow

User: "I need images for a luxury hotel website"

Your response: "I'll help you create prompts for luxury hotel imagery. Let me craft several options:

Lobby shot:

luxury hotel lobby, marble floors, crystal chandelier, modern furniture, warm ambient lighting,
professional architectural photography, wide angle, sophisticated atmosphere --ar 16:9 --style raw --v 7

Suite bedroom:

luxury hotel suite bedroom, king size bed with premium linens, floor to ceiling windows with city view,
elegant minimalist design, soft natural lighting, architectural photography --ar 16:9 --style raw --v 7

Restaurant/dining:

upscale hotel restaurant, elegant table settings, mood lighting, sophisticated ambiance,
fine dining atmosphere, professional interior photography --ar 16:9 --style raw --v 7

These prompts emphasize luxury through specific details (marble, crystal, premium materials) and use architectural photography style with --style raw for photorealism. The 16:9 ratio is ideal for website banners."

---

Notes

• This skill focuses purely on prompt engineering, not API integration
• For actual image generation via Legnext API, use the legnext-midjourney skill
• Always provide the complete prompt the user can copy and use
• Explain your reasoning when helpful (why certain parameters, why specific wording)
• Be ready to iterate - prompting is a creative process