Design & Creative

30 Stable Diffusion Prompts With Negative Prompts and Weights

30 Stable Diffusion prompts for portraits, products, landscapes and art styles, each with a negative prompt, plus the web UI weight syntax in one card.

Robert Youssef10 min
30 Stable Diffusion Prompts With Negative Prompts and Weights
On this page

Stable Diffusion prompts come in pairs: one prompt for what you want in the picture and a second, negative prompt for what you don't. This page has 30 of those pairs for portraits, products, landscapes, art styles and img2img, each ready to paste into the two prompt boxes. Two short sections come first, on how a prompt is put together and on the weight syntax.

How a Stable Diffusion Prompt Is Built

This Stable Diffusion prompt guide starts with the parts, because the 30 prompts below are all built from the same six, shown here in prompt 1:

Part What it sets In prompt 1
Subject Who or what, and where portrait of an elderly fisherman, weathered skin, knitted wool cap, harbor in the background
Medium Photo, painting, illustration 85mm photo
Style The look of that medium shallow depth of field
Lighting Where the light comes from (soft overcast light:1.2)
Color Palette or grade natural color
Quality Level of finish high detail
Prompt 1 cut into its six parts with a label on each, and the weighted lighting phrase highlighted

The prompts on this page are written as short tags separated by commas, in the style of Stable Diffusion 1.5 and SDXL prompts. If you're on a newer model or a hosted service, check its documentation for the prompt style it expects before you assume tags and weights carry over. Our guide on how prompting changes for image models covers images, video and prompting with a picture as input.

Weights, Emphasis and the 75-Token Limit

Prompt emphasis, which people also search for as prompt strength, tells the model which words matter more. The syntax below comes from the AUTOMATIC1111 Stable Diffusion web UI, whose Features page gives this cheat sheet:

You type What it does
(word) Increases attention to the word by a factor of 1.1
((word)) Increases it by a factor of 1.21, which is 1.1 times 1.1
[word] Decreases attention by a factor of 1.1
(word:1.5) Increases attention by a factor of 1.5

This is a feature of that web UI and of interfaces compatible with it. Other Stable Diffusion services may read parentheses differently, so check before you rely on it.

The same page says Stable Diffusion usually accepts 75 tokens, and that typing past 75 in the web UI raises the limit to 150, then further as you keep typing.

The web UI weight syntax as a four-row card, with a bar underneath showing the prompt limit growing from 75 tokens to 150

Portrait and Character Prompt Examples (1 to 8)

Eight Stable Diffusion prompt examples for people and characters. Each block holds the pair: the first line goes in the prompt box, the second in the negative prompt box.

1. Elderly fisherman

A weathered face in flat harbor light.

Prompt: portrait of an elderly fisherman, weathered skin, knitted wool cap, (soft overcast light:1.2), harbor in the background, 85mm photo, shallow depth of field, natural color, high detail
Negative: blurry, deformed hands, extra fingers, oversaturated, cartoon, watermark, text

2. Studio headshot

Clean and neutral, for a profile picture.

Prompt: studio headshot of a woman in her thirties, short dark hair, charcoal blazer, (soft key light from the left:1.1), plain gray backdrop, 85mm photo, sharp eyes, natural skin texture
Negative: blurry, plastic skin, harsh shadows, oversaturated, jewelry, watermark, text

3. Forest ranger

A full-body fantasy character for concept work.

Prompt: full body character, forest ranger, green hooded cloak, leather bracers, longbow on the back, standing on a mossy rock, misty pine forest, concept art, painterly, muted greens, detailed costume
Negative: cropped, extra limbs, deformed hands, modern clothing, text, watermark, low detail

4. Starship pilot

Rim light and a two-color grade.

Prompt: portrait of a starship pilot, orange flight suit, helmet under one arm, hangar lights behind, (rim light:1.2), cinematic still, teal and orange palette, film grain
Negative: blurry, cartoon, deformed face, extra fingers, text, logo, oversaturated

5. Street portrait at night

Neon reflections and a 35mm look.

Prompt: candid street portrait of a young man, denim jacket, neon signs reflected in a wet street, night, 35mm photo, shallow depth of field, moody color
Negative: daylight, studio backdrop, blurry face, extra fingers, readable text, watermark

6. Fox in a raincoat

A picture-book character.

Prompt: a small fox in a yellow raincoat holding an umbrella, children's book illustration, soft gouache texture, warm palette, simple background
Negative: photorealistic, scary, dark, cluttered background, text, watermark

7. Jazz musician in black and white

Hard side light and grain.

Prompt: black and white portrait of a jazz musician, trumpet in hand, smoke in the air, hard side light, high contrast, 50mm photo, film grain
Negative: color, blurry, deformed hands, extra fingers, flat lighting, text, watermark

8. Character sheet

Three views of one design on a plain background.

Prompt: character sheet, young inventor with goggles and a tool belt, front view, side view, back view, neutral pose, plain white background, clean line art, flat colors
Negative: cropped, different outfits, background scenery, extra limbs, text, watermark

Product and Food Prompts (9 to 14)

Product prompts name the surface, the light and the backdrop. The negatives keep logos and stray text off the object.

9. Ceramic mug

Morning light and steam.

Prompt: ceramic coffee mug on a walnut table, morning window light from the left, steam rising, minimalist kitchen, product photography, crisp focus, neutral tones
Negative: clutter, logo, text, harsh shadows, plastic look, low resolution

10. White sneaker

Hard sun and one sharp shadow.

Prompt: white leather sneaker on a concrete block, (hard sunlight:1.2), sharp shadow, pale blue sky backdrop, product photography, crisp focus, minimal
Negative: logo, text, dirt, blurry, cluttered, low resolution

11. Skincare bottle

Wet stone, droplets and a blank label.

Prompt: amber glass skincare bottle with a blank label, wet stone surface, water droplets, soft diffused light, green leaves in the background, product photography, shallow depth of field
Negative: text on label, logo, fingerprints, harsh glare, cluttered, blurry

12. Wristwatch macro

Brushed steel against black.

Prompt: macro photo of a steel wristwatch, black dial, brushed metal texture, dark background, single softbox reflection, high detail
Negative: brand name, logo, dust, scratches, blurry, overexposed

13. Cheeseburger

Side light on a dark slate.

Prompt: cheeseburger on a wooden board, sesame bun, melted cheddar, fresh lettuce, dark slate background, side light, food photography, shallow depth of field
Negative: plastic look, soggy, messy plate, oversaturated, text, watermark, blurry

14. Breakfast flat lay

A top-down table in warm neutrals.

Prompt: flat lay of croissants and a cup of coffee on a linen cloth, morning light, crumbs, top-down view, food photography, warm neutral tones
Negative: hands, clutter, harsh shadows, artificial color, text, watermark

Landscape and Architecture Prompts (15 to 20)

Place, time of day and lens. The negatives remove people and vehicles from scenes that should be empty.

15. Alpine lake at sunrise

Peaks mirrored in still water.

Prompt: alpine lake at sunrise, mirror reflection of snowy peaks, pine trees in the foreground, (golden light:1.2), light mist, wide angle landscape photo, high detail
Negative: people, buildings, oversaturated, blurry, lens flare, text, watermark

16. Dunes at dusk

Long shadows and two colors.

Prompt: sand dunes at dusk, long shadows, rippled texture, deep orange and violet sky, minimal composition, telephoto landscape photo
Negative: people, vehicles, footprints, oversaturated, noise, text, watermark

17. Forest path in fog

Muted color and soft light.

Prompt: forest path in thick fog, tall beech trees, autumn leaves on the ground, soft diffused light, muted colors, 35mm photo, atmospheric
Negative: people, sunbeams, oversaturated, blurry, text, watermark

18. Hillside house

Blue hour and straight verticals.

Prompt: modern concrete house on a hillside, floor-to-ceiling windows, warm interior lights, blue hour, architectural photography, straight verticals, wide angle
Negative: distorted perspective, people, cars, clutter, oversaturated, text, watermark

19. Reading nook

A small interior on a rainy day.

Prompt: small reading nook, armchair by a window, bookshelves, floor lamp, rainy day light, interior photography, warm neutral palette, soft shadows
Negative: people, clutter, distorted furniture, harsh light, text, watermark

20. City intersection from above

Light trails on wet asphalt.

Prompt: aerial view of a city intersection at night, light trails, wet asphalt, grid of streets, high contrast, cool blue tones, long exposure photo
Negative: daylight, blurry, tilted horizon, text, watermark, low resolution

Art Style Prompts (21 to 26)

Six media, each described by its materials and marks, with no artist names. The negatives push away the photo look.

21. Watercolor

Loose washes and white paper.

Prompt: watercolor painting of a harbor town, loose brushwork, wet-on-wet washes, soft edges, pale blue and ochre palette, white paper showing through
Negative: photorealistic, hard outlines, digital look, oversaturated, text, signature

22. Ink drawing

Pen lines and cross-hatching.

Prompt: black ink drawing of an old oak tree, cross-hatching, fine pen lines, white background, high contrast, hand-drawn
Negative: color, gray wash, photorealistic, blurry, text, signature

23. Oil painting

Thick paint and visible canvas.

Prompt: oil painting of a still life with lemons and a blue jug, thick impasto brushstrokes, warm side light, dark background, visible canvas texture
Negative: photorealistic, flat color, digital look, text, signature, frame

24. Anime

Cel shading under a summer sky.

Prompt: anime style illustration, a cyclist on a seaside road, summer afternoon, cel shading, clean line art, bright sky, soft clouds
Negative: photorealistic, 3d render, extra limbs, deformed hands, text, watermark

25. Pixel art

A limited palette and crisp pixels.

Prompt: pixel art, small wizard tower on a hill at night, 16-bit style, limited palette, crisp pixels, starry sky
Negative: blurry, anti-aliasing, gradients, photorealistic, text, watermark

26. Isometric

A cutaway building in pastel.

Prompt: isometric illustration of a tiny coffee shop, cutaway view, pastel colors, clean vector style, soft shadows, plain background
Negative: perspective distortion, photorealistic, clutter, signs with words, text, watermark

img2img Prompts (27 to 30)

These img2img prompt examples start from a picture you supply, so the prompt describes the result you want and leaves out what's already there. The denoising setting decides how much of the original survives: lower stays closer to your picture, higher hands more of it to the prompt.

27. Sketch to watercolor

Upload a line sketch.

Prompt: watercolor painting, loose washes, soft edges, pale palette, white paper texture
Negative: photorealistic, hard outlines, text, signature

28. Photo to flat illustration

Upload a photo of an object or a room.

Prompt: flat vector illustration, clean shapes, limited palette, soft shadows, plain background
Negative: photorealistic, noise, texture, text, watermark

29. Change the season

Upload an outdoor photo.

Prompt: same scene in winter, snow on the ground and rooftops, overcast sky, cool tones, bare trees
Negative: green leaves, summer light, added people, text, watermark

30. Relight a product shot

Upload a product photo taken in poor light.

Prompt: product photography, soft studio light from the left, clean white background, gentle shadow under the object, crisp focus
Negative: harsh glare, colored light, clutter, text, logo, watermark

Negative Prompts That Earn Their Place

The web UI's documentation describes the negative prompt as "another prompt of things the model should avoid." Write it for the picture in front of you. The negatives on this page come from five groups:

  • Anatomy: deformed hands, extra fingers, extra limbs. For people and characters.
  • Quality: blurry, low resolution, noise.
  • Wrong medium: cartoon for a photo, photorealistic for a painting.
  • Unwanted marks: text, watermark, logo, signature.
  • Composition: cropped, cluttered, distorted perspective.

List what would ruin this image and stop there. For how each model handles negatives, there's a full guide to negative prompts.

Two columns side by side: the prompt lists what belongs in the picture, the negative prompt is a separate list of things to keep out

Frequently Asked Questions

What is the best Stable Diffusion prompt format?

For tag-style models, a comma-separated list in a steady order: subject first, then medium, style, lighting, color and quality words, with the negative prompt in its own box. Keeping the order fixed means you can change one part and compare results. The right format still depends on the model you run, so read its documentation before you settle on one.

How do weights like (word:1.5) work in Stable Diffusion?

In the AUTOMATIC1111 web UI, the number after the colon multiplies the model's attention to the words inside the parentheses. Values below 1 turn attention down: the documentation's example (word:0.25) decreases it by a factor of 4. Numbers work with parentheses only, and square brackets take no number. To type literal parentheses in a prompt, put a backslash before each one.

How long can a Stable Diffusion prompt be?

The web UI's documentation gives 75 tokens as the standard and says a longer prompt is split into chunks of 75 tokens that are processed separately and joined. It also describes a BREAK keyword, in capitals, that ends the current chunk so the text after it starts a new one. Other interfaces handle long prompts in their own ways.

Are there Stable Diffusion prompts on GitHub or in a PDF?

Yes, there are public prompt galleries on GitHub. For a PDF, save this page from your browser's print dialog and you'll have the 30 pairs with the syntax card. The library also has a Flux prompt collection if you use that model too.

Your Move

Paste pair 1, generate, then change only the subject and generate again. That shows you what the other five parts are doing. If you'd rather describe a picture in plain words and get a structured prompt back, use our prompt generator. And if you're still choosing a tool, we've compared Midjourney vs Stable Diffusion and listed other AI image generators.

Put this into practice: browse the concept art prompts in the God of Prompt library. Want all of it in one place? The Complete AI Bundle puts the whole prompt library, Custom GPTs and n8n workflow templates behind one payment, with lifetime access.

Prompts for this topic

Put this article to work with ready-to-use prompts from the God of Prompt library.

Keep reading

The best of the blog, in your inbox

One email when notable prompts, tools, and model updates land. No spam, unsubscribe anytime.

Join 100,000+ subscribers. One email a week, real prompts, tools, and model updates. Unsubscribe anytime.