Midjourney & DALL-E Prompting: A Visual Guide

This article is written in English. Our training modules are available in French.

Text-to-image AI has exploded in popularity. But there's an art to writing prompts that produce stunning visuals instead of generic or chaotic images.

Let's explore the fundamentals of image prompting.

How Image Prompts Differ from Text Prompts

When prompting for text, you're having a conversation. When prompting for images, you're describing a scene for an AI to visualize.

Text Prompt (ChatGPT)

Explain how photosynthesis works

Image Prompt (Midjourney/DALL-E)

A peaceful forest clearing at golden hour, sunlight streaming 
through leaves, photorealistic, 8K detail, National Geographic style

Image prompts are descriptive, not conversational.

The Anatomy of an Image Prompt

Effective image prompts typically include:

1. Subject

What's the main focus?

"A lone astronaut on a red desert planet"

2. Style

What aesthetic are you going for?

"oil painting style, impressionist, Monet-inspired"

3. Details

Specifics that enrich the image:

"wearing a weathered spacesuit, dust on visor"

4. Mood/Atmosphere

The feeling you want:

"melancholic, vast emptiness, isolation"

5. Technical Specifications

Quality and format cues:

"8K resolution, cinematic lighting, shallow depth of field"

A Prompt Formula

Here's a simple formula to start:

[Subject] + [Style] + [Details] + [Mood] + [Technical]

Example

A wise old owl perched on an ancient book, 
fantasy illustration style, intricate feather details, 
mystical atmosphere with glowing runes, 
4K digital art, dramatic lighting

Midjourney vs. DALL-E

While both generate images from text, they have different strengths:

| Aspect | Midjourney | DALL-E | |--------|------------|--------| | Style | Artistic, stylized | Photorealistic | | Parameters | --ar, --v, --stylize | Direct size/style options | | Strength | Aesthetics, creativity | Accuracy, prompt following | | Access | Discord-based | API and ChatGPT |

Knowing your tool helps you craft better prompts for it.

Keywords That Transform Images

Certain words dramatically impact output:

Quality Enhancers

• →"8K", "ultra-detailed", "high resolution"
• →"professional photography", "award-winning"
• →"masterpiece", "trending on ArtStation"

Style Keywords

• →"cyberpunk", "steampunk", "art deco"
• →"minimalist", "maximalist", "baroque"
• →"watercolor", "oil painting", "digital art"

Lighting

• →"golden hour", "blue hour", "neon lighting"
• →"dramatic shadows", "soft diffused light"
• →"rim lighting", "volumetric lighting"

Camera/Perspective

• →"aerial view", "macro shot", "wide angle"
• →"shallow depth of field", "bokeh"
• →"fisheye lens", "35mm photograph"

Common Mistakes to Avoid

1. Being Too Vague

❌ "A nice picture of a cat" ✅ "A fluffy Persian cat sleeping on a velvet cushion, golden sunlight, cozy atmosphere"

2. Conflicting Instructions

❌ "Dark and bright, simple and complex" ✅ Pick a consistent direction

3. Too Many Subjects

❌ "A dragon fighting a robot while a wizard watches and there's a spaceship and also a beach" ✅ Focus on one main subject

4. Ignoring Negative Prompts

Many tools let you specify what NOT to include. Use this to avoid common artifacts.

The Iterative Process

Great images rarely come from the first prompt. The process is:

1. →Generate with your initial prompt
2. →Analyze what worked and what didn't
3. →Refine the prompt based on observations
4. →Repeat until you get what you want

Think of it as a conversation with the AI through images.

Key Takeaways

1. →Image prompts are descriptive, not conversational
2. →Include subject, style, details, mood, and technical specs
3. →Specific keywords dramatically impact output quality
4. →Different tools have different strengths
5. →Great results require iteration and refinement

Ready to Master Visual AI?

This article covered the what and why of image prompting. But creating consistent, branded visual content requires deeper techniques.

In our Module 7 — Multimodal & Creative Prompting, you'll learn:

• →Advanced techniques for consistent character design
• →Creating brand-aligned visual content
• →Working with style references and image-to-image
• →Combining text and image AI in workflows
• →Prompt engineering for video generation

→ Explore Module 7: Multimodal & Creative Prompting