AI Image Prompt Generator

Why structured image prompts work better

Image models respond to visual vocabulary. "A fisherman" leaves almost everything to chance: style, light, angle, mood and framing. Each extra element you specify removes a random choice, which is why this builder splits the prompt into subject, style, lighting, camera, composition, mood, colors and aspect ratio. Those are the same decisions a photographer or illustrator makes before starting work.

Order matters too. Most image models give more weight to what comes first, so the subject leads and the details follow. The generator then adapts the syntax to your tool:

• Midjourney gets a compact, comma-separated prompt with the --ar parameter for aspect ratio and --no for things to exclude.
• Stable Diffusion gets separate positive and negative prompts plus a suggested pixel size for the aspect ratio, because most Stable Diffusion interfaces have a dedicated negative prompt field.
• DALL·E and ChatGPT images get a natural-language description, which these models follow more reliably than keyword lists.

Lighting and camera presets are written in photographic terms ("85mm lens, shallow depth of field") because models have seen millions of captioned photos described that way.

Worked example

Say you want a warm, documentary-style photo of a fisherman for a blog header. These are the inputs:

Image tool

Midjourney

Subject

an elderly fisherman mending a net on a wooden pier

Style

photorealistic

Aspect ratio

3:2

Lighting

golden hour sunlight

Camera / shot

eye-level medium shot, 50mm lens

Composition and setting

fisherman on the left third, misty harbour in the background

Medium or texture

35mm film grain

Mood

calm and nostalgic

Colors

muted blues and warm orange

Avoid

text

The builder turns these inputs into the following prompt. Load it into the form above with the Midjourney photo example chip and change anything you like.

an elderly fisherman mending a net on a wooden pier, fisherman on the left third, misty harbour in the background, photorealistic, 35mm film grain, golden hour sunlight, eye-level medium shot, 50mm lens, calm and nostalgic mood, muted blues and warm orange color palette --ar 3:2 --no text

The same inputs with Stable Diffusion selected as the tool produce a positive and a negative prompt instead:

Positive prompt:
an elderly fisherman mending a net on a wooden pier, fisherman on the left third, misty harbour in the background, photorealistic, 35mm film grain, golden hour sunlight, eye-level medium shot, 50mm lens, calm and nostalgic mood, muted blues and warm orange color palette, highly detailed, sharp focus

Negative prompt:
text, blurry, low quality, distorted, extra limbs, watermark, jpeg artifacts

Size: 1216x832 (3:2)

This page shows only the prompt. The answer depends on the AI tool you paste it into, so always review the result before you use it.

Tips to improve your results

• Describe one clear subject. Several subjects in one prompt often blend together.
• Change one variable at a time (only lighting, only camera) to learn what each does.
• Put unwanted elements in the avoid field rather than writing "no text" in the subject.
• Text inside images is still unreliable in many models; add text in an editor afterwards.
• Check the license and usage terms of your image tool before using results commercially.