Prompt · Image generation

Brand image style guide for consistent AI-generated visuals

Create a written visual style guide for a set of generated images: palette, lighting, composition, subjects, words to always and never use, a reusable base prompt and a review checklist.

The prompt

You are an art director creating a visual style guide for images that will be produced with an AI image generator.

Brand or project: [NAME AND ONE-LINE DESCRIPTION]
Personality in three adjectives: [e.g. "calm, practical, optimistic"]
Audience: [WHO WILL SEE THE IMAGES]
Where images will appear: [e.g. "blog headers 16:9", "Instagram 4:5", "app onboarding screens"]
Brand colors (hex if available): [COLORS]
Reference descriptions: [DESCRIBE 2-4 IMAGES OR STYLES YOU LIKE AND WHY, without naming living artists]
Things to avoid: [e.g. "stock-photo handshakes", "neon", "text inside images"]
Image generator: [TOOL NAME]

Produce a style guide with:
1. Visual direction in one paragraph.
2. Fixed style attributes: medium (photo, flat illustration, 3D...), palette, lighting, camera or perspective, level of detail, texture, mood.
3. Subject rules: how people, objects and backgrounds should look; diversity and representation guidelines.
4. Composition rules per format, including safe space for headlines.
5. A base prompt with a [SUBJECT] slot that encodes all fixed attributes, adapted to the syntax of my image generator.
6. A negative list of words or elements to exclude.
7. Five example prompts for different subjects using the base prompt.
8. A review checklist to accept or reject each generated image.

Example output

Example (excerpt):
Visual direction: soft, flat illustrations with generous white space, warm daylight tones and simple geometric shapes; friendly but never childish.
Fixed attributes: flat vector illustration, palette of #1F4E79, #F2B134 and off-white, soft shadows only, no gradients, front or three-quarter view.
Base prompt: "Flat vector illustration of [SUBJECT], limited palette of deep blue, warm yellow and off-white, soft shadows, plenty of empty space on the left, clean geometric shapes, calm optimistic mood"
Checklist: palette matches; no text or logos; hands and faces look natural; left third free for a headline.

When to use it

Use this prompt when you will produce many images over time for the same website, newsletter, course or product and want them to look like a family rather than a random collection. It is also useful when several people generate images and need a shared reference.

The guide itself is written by a text model such as ChatGPT, Claude or Gemini; the base prompt and examples are then used in whichever image generator you prefer. Before you start, collect a handful of images you like and note what they have in common, such as soft daylight, muted colors or lots of empty space. Those observations make far better inputs than a vague request for something “modern” or “clean”.

How to customize it

[PERSONALITY] in three adjectives is easier to translate into visuals than a long brand description.
[WHERE IMAGES WILL APPEAR] sets aspect ratios and how much space must be left for text.
[BRAND COLORS] work best as hex codes, though image models follow color names more reliably than codes; keep both.
[REFERENCE DESCRIPTIONS] should describe qualities (lighting, texture, mood) rather than imitate a specific living artist.
[IMAGE GENERATOR] matters because prompt syntax, negative prompts and parameters differ between tools.

Why it works

Consistency comes from fixing most variables and changing only one. The guide separates fixed attributes from the [SUBJECT] slot, so each new prompt changes the content but not the look. Writing it down also makes the style independent of a single person’s memory.

The negative list and the review checklist handle the other half of consistency: deciding what to reject. Generators produce variation even with identical prompts, so a short acceptance checklist saves time and keeps the set coherent.

Tips and variations

Generate a test batch with the five example prompts and revise the guide based on what goes wrong.
If your tool supports style or image references, save your best results and reuse them.
For recurring people or mascots, add our consistent character prompt.
Learn the building blocks in the image generation prompt template and our guide on choosing an image generator.
Check each tool’s licensing terms before commercial use.

#Image Generation #Brand Guidelines #Art Direction #Visual Consistency

More Image generation prompts

View all

Image generationadvanced

Keep a character consistent across multiple images

Step 1 - Character sheet (write once, reuse unchanged): CHARACTER: [NAME], a [AGE/TYPE, e.g. "small fox mascot", "woman in her 60s"] with [FACE AND BODY: 3-5 fixed, specific features, e.g. "round amber eyes, white tip on the left ear only, short stocky build"], wearing [OUTFIT: fixed items and colours, e.g. "a mustard-yellow knitted scarf and a teal backpack"], drawn in [STYLE, e.g. "soft 2D children's book illustration, thick clean outlines, flat colours with light grain"]. Step 2 - Reference image: [CHARACTER SHEET], character turnaround: front view, side view and back view, neutral pose, plain light grey background, full body, even lighting. Step 3 - Scene template (one per image): [CHARACTER SHEET], [ACTION], in [SETTING], [TIME OF DAY AND LIGHT], [SHOT TYPE AND ANGLE]. Keep: same face, same markings, same outfit and colours, same art style. Avoid: [e.g. "different outfit, extra accessories, realistic fur, text"]. Aspect ratio: [RATIO] Tool settings (use what your tool supports): - Attach the Step 2 image as a character or style reference. - Reuse the same seed while testing to compare changes. - Keep the model version and style settings identical for the whole series.

Works with: Midjourney, FLUX, Stable Diffusion, GPT Image, Imagen

Image generationbeginner

Image generation prompt template: subject, style, light, composition

[SUBJECT: who or what, with 2-3 specific details], [ACTION OR POSE], in [SETTING: place, time of day, season, weather]. Style: [e.g. "documentary photograph", "watercolour illustration", "flat vector icon", "3D clay render"], inspired by [GENRE OR ERA, e.g. "1970s travel posters"]. Lighting: [e.g. "soft golden-hour backlight", "overcast diffused light", "neon rim light", "harsh midday sun"]. Camera and composition: [SHOT TYPE, e.g. "close-up", "wide establishing shot"], [ANGLE, e.g. "eye level", "low angle", "top-down"], [LENS FEEL, e.g. "35mm, shallow depth of field"], subject placed [e.g. "on the left third, empty space on the right for text"]. Colour palette: [2-4 colours or a mood, e.g. "muted teal and warm orange"]. Mood: [ONE OR TWO WORDS, e.g. "calm, hopeful"]. Avoid: [UNWANTED ELEMENTS, e.g. "text, logos, watermarks, extra fingers, clutter"]. Aspect ratio: [e.g. 16:9 for banners, 1:1 for social posts, 9:16 for stories, 2:3 for posters]

Works with: Midjourney, FLUX, Stable Diffusion, DALL-E, Imagen

Image generationintermediate

Infographic layout plan with accurate data and visual prompts

You are an information designer. Plan an infographic from the content below. The layout will be built in a design tool; AI image generation will be used only for illustrations and backgrounds. Topic: [TOPIC] Audience and where it will be shared: [e.g. "parents, shared on Instagram and printed as an A4 poster"] Format and size: [e.g. "1080x1350 px portrait", "A4 portrait"] Brand colors and fonts (if any): [DETAILS] Facts and numbers (with sources): """ [PASTE DATA POINTS AND WHERE EACH COMES FROM] """ Produce: 1. A headline (under 10 words) and a subheadline that states the main takeaway. 2. A layout plan from top to bottom: sections, approximate share of the height, reading order. 3. For each data point: the best visual form (bar, icon array, big number, timeline, comparison) and why; the exact text and number to show. 4. All on-image text in final wording, kept short, with the source line for the footer. 5. Prompts for illustrations or icons in one consistent style, with transparent or plain backgrounds and no text inside the images. 6. An accessibility note: color contrast, minimum font sizes, and alt text for the whole infographic. 7. A data check: list every number as it will appear next to the source value, so I can verify nothing was rounded or changed incorrectly. Do not add facts that are not in my list.

Works with: ChatGPT, Claude, Gemini, Any image model