How to Structure Prompts

Before defining your subject or scene, establish the visual rules of the image.

This opening block is what shapes consistency across a campaign and removes randomness from the output.

This is where you outline:

• Camera Technique
• Lighting Behaviour
• Lens Physics
• Imperfections
• Textural Mood

It sets the tone before anything else appears in the frame.

This is the foundation of controlled prompting.

This section defines how the image should feel. Not the subject. Not the props. The behaviour.

Think of it as describing the image like a photographer would describe a lighting setup or lens characteristic.

This might include:

• Tight or mid framing
• Shallow depth of field
• Handheld or tripod stability
• Halation in high-contrast areas
• Minor imperfections that give realism

This ensures the image behaves with intention rather than defaulting to generic, inconsistent styling.

This is the cleanest and most literal component of your structure. One concise line.

Formula:

Object + Action/State + Key detail

Keep it objective. Avoid unnecessary adjectives. The subject line anchors the entire composition.

Examples:

• Subject: A chrome perfume bottle standing upright, soft condensation forming
• Subject: A studio headshot of a male model in three quarter profile, defined jawline catching a crisp rim light
• Subject: A studio portrait of a female model facing camera, clean mid framing, soft highlight rolling across her cheekbone

A concise subject anchors the shot, preventing drift
and keeping the composition controlled.

Next comes the world the object sits in. This is where you build depth, tone, and spatial logic.

Scene Structure:

Surface + Backdrop + Set elements + Atmospheric detail

Examples:

• Scene: Resting on matte stone with acrylic blocks positioned behind, faint haze catching the backlight
• Scene: Set on brushed metal with dark geometric shadows and controlled reflection
• Scene: Draped over soft ivory fabric with gentle folds and a warm atmospheric glow

This line tells the model how the space behaves around the subject.

Tags reinforce everything above. They stabilise consistency, especially across a full series or campaign.

Tags should be minimal, technical, and high-signal.

Examples:

• Editorial lighting
• Macro realism
• 35mm texture
• Clean controlled lighting
• Studio still life

Tags function like the glue that holds the style together.

Here is the exact four-part structure you use:

Camera + Aesthetic Block:
[Camera technique] [Lens physics]
[Lighting behaviour] [Imperfections] [Mood]

Subject:
[Object] [Action/State] [Key detail]

Scene:
[Surface] + [Backdrop] + [Set elements] + [Atmospheric detail]

Tags:
[Style anchors] [Technical cues] [Film/texture terms]

Even strong prompts fall apart when the structure is inconsistent. These are the errors that interrupt clarity, confuse the model, and produce images that feel generic or unstable. Use this page as a reference for what to avoid.

• Starting with the subject: Leads to generic results. Define camera behaviour and lighting first so the image has a stylistic foundation.
• Vague lighting descriptions: “Soft lighting” isn’t enough. Specify direction, behaviour, and how it interacts with surfaces.
• Skipping technical tags: Tags stabilise style across multiple images. Without them, outputs drift inconsistently.warm atmospheric glow
• Writing everything in one sentence: Blended prompts reduce clarity. Keep the four parts separate for clean hierarchy.
• Changing too many variables at once: When testing, adjust one element at a time. This is how you refine consistency.

The goal isn’t complexity, but control. Build your prompts with intention, and the quality follows.

Use this structure to write clearer, more consistent prompts.