AI Prompts

Midjourney prompts.

Ready-to-use prompt formulas for photoreal, design, art and full parameter control — tap to copy, then swap in your subject.

Tap Copy on any prompt, paste it into your AI assistant, and replace the [bracketed] parts with your own details.

Photorealistic

Portrait

Portrait of a [age] [subject], [expression], [wardrobe], soft natural window light, shallow depth of field, shot on an 85mm lens at f/1.8, detailed skin texture --ar 4:5 --style raw

Scene / street

[Subject] in [location] at [time of day], [weather or mood], cinematic lighting, reflections, shot on 35mm film, atmospheric --ar 16:9 --style raw

Product photo

Commercial product photo of [product] on [surface], soft directional light from the [side], subtle shadows, sharp focus, minimalist background --ar 1:1

Design & product

Logo / vector

Minimalist vector logo of a [subject], geometric, [two colors], flat design, clean lines, no shading, white background --no text photorealism --ar 1:1

Image models can over-add text and detail; the --no parameter helps keep logos clean.

UI mockup

High-fidelity UI mockup of a [app type] screen, clean [light/dark] interface, [accent color] highlights, modern sans-serif type, generous spacing, Dribbble aesthetic --ar 3:2

Icon set

A set of [style, e.g. rounded line] icons for [theme], consistent stroke weight, single accent color, centered on white, simple and legible at small sizes --ar 1:1

Art & illustration

Illustration

[Subject and scene] in [art style, e.g. watercolor / flat editorial / storybook], [color palette], [mood], hand-crafted texture --ar 3:2

Anime style

[Subject] in [setting], soft hand-painted anime style, gentle pastel colors, detailed background, warm atmosphere --niji 6 --ar 16:9

Concept environment

Concept art of [place], [time of day] lighting, epic scale, atmospheric depth, painterly detail, [mood] --ar 21:9

Parameters & control

Set the shape

[your prompt] --ar 16:9

--ar sets aspect ratio: 1:1 square, 4:5 portrait, 16:9 wide, 21:9 cinematic.

Control the style strength

[your prompt] --stylize 250

Lower values follow your words more literally; higher (up to 1000) give more artistic license. Pair a low value with --style raw for realism.

Match a reference image

[your prompt] --sref [image URL] --sw 100

--sref matches the colour and look of a reference; --sw sets how strongly, from 0 to 1000.

Exclude unwanted things

[your prompt] --no [things to avoid]

Describe what you DO want in the prompt and use --no to remove elements, e.g. --no text, blur.

Prompting images is different

Unlike a chatbot, Midjourney doesn't need polite sentences — it responds to a clear formula: subject, scene, lighting, medium or camera, style, then parameters at the end. Start simple, generate four, then refine the best one. Note that Midjourney's parameters change between versions, so check the current parameter list if a flag doesn't behave. For text-to-image with other tools, see the general image prompts.

FAQ

Do these work in the latest Midjourney version?
The descriptive parts work across versions; the parameters (like --ar, --stylize and --sref) occasionally change syntax between V6, V7 and V8. If a flag doesn't behave, check Midjourney's current parameter list.
Why avoid saying 'no X' in the description?
Image models can latch onto words you mention even when you say 'no'. Describe what you DO want, and use the --no parameter to exclude things instead.

More prompt packs

Image prompts →
Midjourney, DALL·E & SD templates
Video prompts →
Sora, Veo & Runway templates
Social media prompts →
Captions, hooks & content ideas
Writing prompts →
Email, rewriting, summaries & ideas