Prompt Framework

Midjourney Prompt Framework: Structure Visual Prompts for Consistent Results

Midjourney output quality is almost entirely determined by prompt structure. The difference between a generic, inconsistent image and a polished, production-ready result comes down to how specifically and in what order you describe the visual elements. This framework gives you a repeatable structure that produces reliable results across use cases.

How Midjourney reads prompts

Understanding how Midjourney processes text helps you write better prompts:

Word order matters. Words and phrases at the beginning of a prompt carry more weight than those at the end. The most important visual element should come first.
Concrete visual vocabulary outperforms abstract descriptions. 'Mysterious and atmospheric' produces unpredictable results. 'Dramatic fog, long shadows, blue-hour lighting' produces specific, replicable visual output.
Commas separate concepts; double colons (::) separate elements with weighted importance. Advanced users use :: to weight specific elements: 'forest::2 cabin::1' gives the forest twice the visual weight.
Negative prompts (--no) are often more effective than positive constraints. Adding '--no text' or '--no people' is cleaner than trying to describe the absence of something.

The Midjourney framework: 7 elements

1. Subject

The primary focus of the image -- described specifically. Include: what it is, its most important visual characteristics, and any relevant action or state. The more specific, the more predictable the output.

Weak: "a dog" | Strong: "a golden retriever sitting alertly in tall grass, facing camera, ears forward"

2. Style / Art Direction

The overall visual style: photography, illustration, painting style, or specific art movement. This is the single most impactful element after the subject for controlling the aesthetic.

Examples: commercial product photography, oil painting in the style of Dutch Golden Age, vector illustration, cinematic still, architectural rendering, watercolor illustration

3. Medium

The physical or digital medium -- relevant for illustration and art styles. Affects texture, finish, and aesthetic quality in the output.

Examples: oil on canvas, gouache, digital matte painting, ink wash, charcoal drawing, vector art, 3D render

4. Lighting

Lighting is one of the most powerful elements for controlling mood and quality. Professional-sounding lighting terms consistently produce better output than vague mood descriptions.

Examples: golden hour, blue hour, studio softbox, rim lighting, Rembrandt lighting, natural window light from the left, dramatic side lighting, overcast diffused, neon glow

5. Composition / Camera

Camera angle, perspective, and framing. For photographic prompts, camera and lens specifications significantly improve realism.

Examples: close-up portrait, overhead flat lay, wide establishing shot, 35mm eye-level, 85mm f/1.8 portrait lens, rule of thirds, centered symmetrical, bird's-eye view

6. Mood / Atmosphere

The emotional quality of the image -- described with visual language rather than abstract emotions.

Weak: "sad and lonely" | Strong: "muted desaturated tones, empty background, long shadows, quiet and still"

7. Parameters

Technical flags that control output dimensions, style strength, and model version.

--ar [aspect ratio]   : 16:9 (landscape), 4:5 (portrait), 1:1 (square), 9:16 (vertical)
--v 6.1              : use the current model version for best quality
--style raw          : less stylization, more literal/photographic
--no [element]       : exclude something (--no text, --no people, --no watermark)
--q 2                : higher quality rendering (slower)

The complete framework formula

[Subject with specific details], [art direction/style], [medium if relevant],
[lighting description], [composition/camera], [mood through visual language],
[color palette or tone if important] --ar [ratio] --v 6.1 [--style raw if photographic]

Framework applied: 6 real examples

Commercial product photography

Minimalist glass perfume bottle with silver cap on dark polished marble surface, drops of water on the bottle, product photography, studio softbox lighting with subtle rim light, 50mm commercial lens, shallow depth of field, moody luxury aesthetic, dark background, premium fragrance brand feel --ar 4:5 --v 6.1 --style raw

Brand lifestyle image

Young woman in her late 20s working at a tidy white desk with plants, warm morning light from a nearby window, lifestyle photography, natural light, authentic candid moment, soft warm tones, modern home office, remote work aesthetic --ar 16:9 --v 6.1

Editorial illustration

A business professional navigating a complex maze made of financial charts and graphs, digital editorial illustration, flat design with depth, corporate blue and gold palette, isometric perspective, conceptual business metaphor, magazine cover quality --ar 4:5 --v 6.1

Character concept art

A weather-worn female ranger in her 40s wearing layered leather and wool, holding a carved walking staff, standing at a misty mountain pass, fantasy concept art, painterly illustration, warm lantern light against cold blue fog, determined expression, realistic character design --ar 2:3 --v 6.1

Architectural photography

Modern Scandinavian cabin interior with floor-to-ceiling windows overlooking a pine forest, late afternoon golden light filling the space, architectural interior photography, 24mm wide angle lens, warm wood tones, minimal furniture, cozy and editorial --ar 16:9 --v 6.1 --style raw

Social media flat lay

Overhead flat lay of morning coffee setup: ceramic white mug, French press, small vase of eucalyptus, open journal with pen, white linen surface, soft diffused natural light from the left, minimal and clean, Instagram lifestyle photography, warm neutral tones --ar 1:1 --v 6.1

Advanced techniques: style consistency across images

For projects requiring visual consistency across multiple images (brand identity, product catalog, campaign), these techniques help maintain a coherent look:

Use identical lighting and camera terms across all prompts. Keeping 'studio softbox lighting, 50mm lens' consistent across a product series produces a coherent catalog aesthetic.
Define a fixed color palette description. 'Muted sage green, warm cream, dusty terracotta' as a recurring element creates visual brand cohesion.
Use Midjourney's --sref flag (style reference). Reference a seed image URL with --sref to maintain consistent style across generations.
Keep --ar consistent. All assets in a series should use the same aspect ratio to work together in layouts.

Common Midjourney framework mistakes

Starting with style instead of subject. The subject should come first -- it's the anchor of the image. Style, lighting, and composition describe how to render the subject, not replace it.
Abstract emotional language instead of visual language. 'Sad,' 'happy,' 'mysterious' are interpreted inconsistently. Describe the visual elements that create those feelings: 'muted grays, soft focus, downcast expression.'
Not specifying aspect ratio. Default output is square. Always add --ar to match your intended use: 4:5 for Instagram portrait, 16:9 for banner, 9:16 for Stories.
Over-long prompts that dilute focus. More than 25–30 words often reduces output quality as the model tries to balance too many competing visual elements. Prioritize ruthlessly.
Not using --style raw for photographic outputs. The default stylization adds an artistic quality. For product photography, portraits, and realistic imagery, --style raw consistently produces more literal, usable results.