Beginner-Friendly Stable Diffusion Prompt Examples

Tested prompts for stable diffusion prompt examples for beginners compared across 5 leading AI models.

BEST BY JUDGE SCORE Claude Opus 4.7 8/10

If you just installed Stable Diffusion and typed something like 'a dog' and got a blurry mess, you are not doing it wrong. You are just missing the prompt structure that makes the model produce sharp, intentional images. Stable Diffusion responds to specific language patterns, and without knowing those patterns, even great ideas produce mediocre results.

The core problem beginners face is that Stable Diffusion is not a search engine. It does not infer what you mean. It reads every word as a weighted instruction. A prompt like 'beautiful landscape' gives the model almost nothing to work with. A prompt like 'rolling green hills at golden hour, soft volumetric light, 8k, photorealistic, shot on Sony A7R IV' gives it a clear target.

This page walks you through real prompt examples built for beginners, covering the structure, the vocabulary, and the logic behind what works. You will see exactly how prompts are written, why certain modifiers matter, and which mistakes to stop making immediately. Every example here is tested and reproducible on base Stable Diffusion models.

When to use this

Stable Diffusion prompt examples are most useful when you are learning the vocabulary of the model, trying to reproduce a specific visual style, or troubleshooting why your outputs look generic. If you have a clear image in your head and keep getting something different, studying working examples is the fastest path to closing that gap.

You are generating concept art or illustrations for a personal or client project and need consistent quality
You want to create a specific photographic style, such as film grain portraits or golden-hour landscapes, from scratch
You are building a content pipeline and need repeatable prompts that produce on-brand visuals
You are learning the model for the first time and want to understand how prompt structure affects output
You are iterating on a character design or scene and need a baseline prompt to modify

When this format breaks down

If you need precise control over specific object placement or exact facial likeness, text prompts alone will not get you there. Use ControlNet, img2img, or inpainting workflows instead.
If your project requires legally defensible originality, such as commercial illustration with unique IP, relying entirely on prompt-driven generation without significant post-editing introduces IP risk.
If you are trying to recreate a specific existing artwork style with an artist's name, many newer fine-tuned models and SDXL have reduced or removed support for named artist styles, making results inconsistent.
If you need photorealistic human portraits with accurate anatomy and no artifacts, base SD models struggle here. Purpose-trained models like Realistic Vision or use of ADetailer are better suited.

The prompt we tested

You are a Stable Diffusion prompt coach for beginners. Given the subject or scene the user wants to create, generate beginner-friendly Stable Diffusion prompt examples that they can copy and paste directly into any Stable Diffusion interface.

Follow these rules: Return 3-5 complete example prompts formatted as a numbered list, where each example includes (a) a short label describing the style (e.g., 'Photorealistic portrait', 'Fantasy illustration'), (b) a ready-to-paste positive prompt using comma-separated tags with subject, style, lighting, and quality boosters, and (c) a matching negative prompt. End with 2-3 short beginner tips explaining which tags to swap to customize the result. Keep language simple and avoid advanced techniques like ControlNet or LoRAs.

User's subject or scene idea:
I'm brand new to Stable Diffusion and I want to make an image of a cute orange cat sitting by a window on a rainy day. I don't know what words or tags to use to get a nice looking result. Can you show me some example prompts I can try?

Now produce the beginner Stable Diffusion prompt examples.

Model comparison

Model	Quality	Cost	Latency	Tokens
Claude Opus 4.7	8	$0.0775	12.7s	390 / 956
Claude Haiku 4.5	7	$0.0029	6.2s	271 / 528
GPT-5.4	8	$0.0106	14.0s	236 / 669
Gemini 2.5 Pro	4	$0.0016	17.2s	236 / 253
Grok 4.1 Fast Reasoning	7	$0.0003	4.2s	380 / 388

Outputs

Claude Opus 4.7 8/10

# Stable Diffusion Prompts: Cute Orange Cat by a Rainy Window

Here are 4 ready-to-copy prompts in different styles. Just paste them into your Stable Diffusion interface!

---

### 1. Photorealistic Photo

**Positive prompt:**
```
cute orange tabby cat sitting on windowsill, looking out window, rainy day, raindrops on glass, cozy atmosphere, soft natural lighting, shallow depth of field, photorealistic, highly detailed, sharp focus, 4k, professional photography
```

**Negative prompt:**
```
cartoon, anime, painting, blurry, low quality, distorted, extra limbs, deformed, ugly, watermark, text, signature
```

---

### 2. Cozy Anime Style

**Positive prompt:**
```
adorable orange cat by a window, rainy day outside, water droplets on glass, anime style, Studio Ghibli inspired, warm cozy lighting, soft colors, detailed illustration, beautiful background, masterpiece, best quality
```

**Negative prompt:**
```
realistic, photograph, 3d render, blurry, low quality, bad anatomy, extra legs, malformed, watermark, text
```

---

### 3. Watercolor Painting

**Positive prompt:**
```
orange kitten sitting on wooden windowsill, gazing at rain, watercolor painting, soft brush strokes, pastel colors, dreamy mood, artistic, hand-painted, delicate details, beautiful composition
```

**Negative prompt:**
```
photograph, realistic, 3d, harsh lines, low quality, blurry, deformed, ugly, extra limbs, signature, watermark
```

---

### 4. Digital Art / Illustration

**Positive prompt:**
```
cute fluffy orange cat, sitting by rainy window, cozy indoor scene, warm lamp light, city lights blurred outside, digital art, concept art, trending on artstation, vibrant colors, detailed fur, cinematic lighting, highly detailed
```

**Negative prompt:**
```
photo, realistic, low quality, blurry, bad proportions, extra tail, deformed face, ugly, text, watermark, signature
```

---

## 💡 Beginner Tips

1. **Swap the cat description** to change the look: try `black cat`, `fluffy white kitten`, or `calico cat` instead of `orange tabby`.

2. **Change the mood by swapping lighting tags**: replace `warm cozy lighting` with `moody lighting`, `golden hour`, or `dramatic lighting` for a different feel.

3. **Change the style** by replacing tags like `photorealistic` with `oil painting`, `pixel art`, or `3d render` — this is the fastest way to get a completely different image from the same scene!

Claude Haiku 4.5 7/10

# Stable Diffusion Prompts: Cute Orange Cat by Rainy Window

## 1. Cozy Illustration Style
**Positive Prompt:**
cute orange tabby cat sitting on windowsill, rainy day outside, rain drops on window, warm indoor lighting, cozy atmosphere, soft watercolor illustration style, gentle colors, highly detailed, artstation

**Negative Prompt:**
blurry, ugly, distorted cat, dark, sad, realistic photo

---

## 2. Photorealistic Portrait
**Positive Prompt:**
adorable fluffy orange cat by window, rain falling outside, water droplets on glass, natural daylight, cozy home interior, professional photography, sharp focus, 4k, detailed fur

**Negative Prompt:**
cartoon, illustration, blurry, dark lighting, deformed cat, multiple cats

---

## 3. Studio Painting Style
**Positive Prompt:**
cute ginger cat looking out rainy window, oil painting style, soft brushstrokes, warm golden hour light, peaceful mood, detailed background, fine art, museum quality, beautiful composition

**Negative Prompt:**
digital art, sketch, low quality, oversaturated, ugly cat face, distorted

---

## 4. Storybook Aesthetic
**Positive Prompt:**
sweet orange kitten sitting windowsill, watching rain, children's book illustration, whimsical, soft pastel colors, cozy room interior, detailed linework, charming, storybook style

**Negative Prompt:**
dark, photorealistic, scary, blurry, deformed, ugly proportions

---

## Beginner Tips:

**Tip 1 – Swap the cat description** to change breeds or style. Try "fluffy Persian orange cat" or "sleek ginger cat" instead of "tabby cat" to get different looks.

**Tip 2 – Change the art style** by replacing "watercolor," "oil painting," or "illustration" with styles you like: "anime," "3D render," "pencil drawing," or "CGI."

**Tip 3 – Adjust mood** by swapping words like "cozy" and "warm" for different feelings: try "moody," "peaceful," "magical," or "melancholic" to change the image's overall vibe.

GPT-5.4 8/10

1. **Soft photorealistic cat**
   - **Positive prompt:** `cute orange cat, sitting by a window, rainy day, raindrops on glass, cozy indoor scene, soft natural light, fluffy fur, detailed eyes, calm expression, photorealistic, warm colors, shallow depth of field, highly detailed, high quality`
   - **Negative prompt:** `blurry, low quality, low resolution, bad anatomy, extra legs, extra tail, deformed face, duplicate cat, cropped, out of frame, text, watermark, logo, oversaturated, noisy`

2. **Cozy storybook illustration**
   - **Positive prompt:** `cute orange cat, sitting on a windowsill, rainy window, cozy room, storybook illustration, soft painting, warm cozy atmosphere, gentle lighting, fluffy fur, adorable, detailed background, soft colors, high quality, charming, whimsical`
   - **Negative prompt:** `blurry, messy, low quality, low detail, extra limbs, deformed cat, scary, aggressive, dark shadows, text, watermark, logo, cropped, out of frame`

3. **Anime-style rainy window scene**
   - **Positive prompt:** `cute orange cat, sitting by a window, rainy day, anime style, soft shading, clean lineart, sparkling eyes, fluffy tail, cozy indoor lighting, raindrops on the window, pastel colors, adorable, highly detailed, high quality`
   - **Negative prompt:** `blurry, low quality, bad proportions, extra ears, extra legs, deformed paws, ugly, realistic human features, text, watermark, logo, cropped, dull colors`

4. **Painterly cozy home scene**
   - **Positive prompt:** `orange tabby cat, sitting beside a window, rainy afternoon, digital painting, painterly brushstrokes, cozy home interior, soft light, reflections on wet glass, relaxed pose, detailed fur, beautiful composition, warm tones, high quality, atmospheric`
   - **Negative prompt:** `blurry, low detail, flat lighting, bad anatomy, extra limbs, duplicate animal, distorted face, cluttered background, text, watermark, logo, oversharpened, noisy`

5. **Dreamy cinematic cat portrait**
   - **Positive prompt:** `cute orange cat, by a rain-covered window, cinematic lighting, dreamy mood, soft glow, fluffy fur, detailed whiskers, reflective glass, cozy and peaceful, close-up portrait, realistic style, rich detail, high quality, beautiful bokeh`
   - **Negative prompt:** `blurry, low quality, low resolution, bad anatomy, extra whiskers, extra eyes, duplicate cat, distorted body, harsh shadows, text, watermark, logo, frame, oversaturated`

**Beginner tips:**
- Swap the **style tags** to change the look: try `photorealistic`, `anime style`, `storybook illustration`, or `digital painting`.
- Change the **lighting and mood** tags for a different feeling: `soft natural light`, `warm cozy lighting`, `cinematic lighting`, or `dreamy glow`.
- To change the cat, replace `cute orange cat` with `orange tabby kitten`, `fluffy ginger cat`, or `sleepy orange cat`.

Gemini 2.5 Pro 4/10

Of course! Welcome to the exciting world of Stable Diffusion. It can seem a little confusing at first, but you'll get the hang of it quickly. The key is to describe what you want, the style you want it in, and add a few words to boost the quality.

Here are a few beginner-friendly examples for your idea of a cute orange cat on a rainy day. You can copy and paste these directly to get started.

***

### 1. Photorealistic Photo

This prompt aims for an image that looks like a real, high-quality photograph.

*   **Prompt:**
    ```
    masterpiece, best quality, photograph of a cute orange tabby cat sitting on a cozy windowsill, sleepy, rainy day, raindrops on the window glass, soft natural light, looking at camera, dslr, sharp focus, high detail
    ```
*   **Negative Prompt:**
    ```
    cartoon, painting, anime, deformed, bad anatomy, disfigured, poorly drawn face, mutated, extra limb, ugly, poorly drawn hands, missing limb, blurry, floating limbs, disconnected limbs, malformed, out of focus, long neck, long body
    ```

### 2. Cozy

Grok 4.1 Fast Reasoning 7/10

1. **Photorealistic scene**  
   Positive: cute orange tabby cat sitting by a window, rainy day outside, soft raindrops on glass, cozy room interior, warm lighting from inside, highly detailed fur, adorable expression, photorealistic, masterpiece, best quality  
   Negative: blurry, low resolution, deformed, ugly, extra limbs, bad anatomy, watermark  

2. **Cute cartoon style**  
   Positive: cute orange tabby cat sitting by a window, rainy day with puddles outside, big sparkling eyes, chibi style, vibrant colors, soft shadows, cartoon illustration, adorable and whimsical, high detail, masterpiece  
   Negative: realistic, photorealistic, blurry, dark, mutated, poorly drawn face, extra fingers  

3. **Cozy painterly art**  
   Positive: cute orange tabby cat sitting by window sill, rainy weather outside, misty rain, warm golden hour light inside, fluffy fur texture, serene atmosphere, oil painting style, detailed brush strokes, high quality, masterpiece  
   Negative: blurry, pixelated, lowres, overexposed, underexposed, distorted proportions  

4. **Fantasy illustration**  
   Positive: adorable orange tabby cat perched by a large window, heavy rain pouring outside, magical glow from candlelight, fantasy art style, intricate details, soft bokeh rain effect, cute and enchanting, ultra detailed, best quality  
   Negative: cartoon, sketch, blurry, grainy, bad composition, text, signature  

**Beginner tips:**  
- Swap "orange tabby cat" with "fluffy white dog" or "siamese kitten" to change the animal.  
- Replace "rainy day" with "sunny afternoon" or "snowy evening" for a different mood or weather.  
- Add "highly detailed eyes" or "dynamic pose" to the positive prompt for more focus on specific features.

What makes these work

01

Start With Subject, Then Context
Your prompt should open with the main subject, then add the setting, then lighting, then style. This order mirrors how the model weights tokens. Starting with style or mood before the subject often causes the subject to be underrepresented in the final image. Write 'a red barn in a snowy field, overcast sky, photorealistic' not 'photorealistic snowy scene with a barn.'
02

Use Quality Boosters Consistently
Terms like 'highly detailed,' '8k,' 'sharp focus,' 'intricate,' and 'trending on artstation' act as quality signals that push the model toward higher-fidelity outputs. They do not guarantee quality on their own, but when combined with a well-structured prompt they consistently improve output sharpness and detail. Place them at the end of your prompt.
03

Name a Visual Style or Medium
Stable Diffusion was trained on labeled images, so naming a medium gives the model a strong anchor. 'Oil painting,' 'watercolor,' 'digital art,' 'photography,' 'concept art,' and 'pencil sketch' each unlock a distinct visual language. Beginners who skip this step get a muddled mix of styles. Picking one keeps the output coherent.
04

Add Lighting as a Dedicated Modifier
Lighting transforms mood more than almost any other variable. Phrases like 'golden hour lighting,' 'soft diffused light,' 'dramatic chiaroscuro,' 'neon backlighting,' and 'studio softbox lighting' each produce dramatically different results on the same subject. Treat lighting as its own prompt clause, not an afterthought.

More example scenarios

#01 · Fantasy character portrait for a game asset

Input

A female elf warrior with silver hair and golden armor, standing in an ancient forest, dramatic rim lighting, detailed face, fantasy art style, highly detailed, 8k, artstation trending, illustration by Greg Rutkowski

Expected output

A sharp, painterly portrait of an elf warrior with detailed facial features, silver hair catching rim light, and intricately textured gold armor. The forest background is atmospheric with depth of field. The output reads as professional concept art suitable for game production reference.

#02 · Photorealistic product mockup for e-commerce

Input

A minimalist white ceramic coffee mug on a marble countertop, soft natural window light, shallow depth of field, commercial product photography, 50mm lens, clean background, ultra-sharp, studio lighting

Expected output

A clean product-style image showing the mug centered on marble, with soft specular highlights and a slightly blurred background. The light source reads as natural and the composition matches standard e-commerce photography conventions.

#03 · Sci-fi environment for a short film storyboard

Input

Interior of a massive abandoned space station corridor, broken windows showing deep space, flickering fluorescent lights, industrial metal panels, volumetric fog, cinematic composition, 16:9, concept art, detailed, moody atmosphere

Expected output

A wide cinematic frame showing a decayed space corridor with strong perspective lines leading to a shattered viewport. Volumetric light rays pass through the fog, metal surfaces show wear and rust, and the color palette leans into cool blues and greens with warm flicker accents.

#04 · Vintage travel poster for a design project

Input

Retro travel poster of Kyoto Japan, cherry blossom trees, Mount Fuji in the background, bold flat colors, art deco typography style, 1950s poster illustration, limited color palette, clean graphic design

Expected output

A flat-color illustration in a 1950s poster aesthetic showing stylized cherry blossoms framing a simplified Mount Fuji. The palette is limited to four or five bold colors and the composition mimics vintage tourism advertising with a clean, graphic quality.

#05 · Children's book illustration for a storybook page

Input

A friendly cartoon fox wearing a red scarf sitting under a big oak tree in autumn, warm soft colors, watercolor style, gentle lighting, storybook illustration, charming and whimsical, suitable for children ages 4-8

Expected output

A warm, softly rendered watercolor-style image of a cartoon fox with large expressive eyes and a bright red scarf. Autumn leaves in oranges and yellows surround the scene. The linework is gentle and the overall tone is cozy and age-appropriate for picture books.

Common mistakes to avoid

Prompts That Are Too Vague
Writing 'a nice sunset' or 'cool robot' gives the model almost no constraint. Vague prompts produce average-looking outputs because the model fills in all the gaps with statistical defaults. Every adjective and detail you add steers the model away from generic and toward specific. Aim for at least 15-25 words in your prompt.
Overloading With Conflicting Styles
Asking for 'photorealistic watercolor oil painting sketch' in one prompt forces the model to blend incompatible aesthetics, resulting in a muddy, undefined output. Pick one primary style and stick to it. If you want a hybrid, use a fine-tuned model built for that specific look rather than stacking contradictory style terms.
Ignoring Negative Prompts
Stable Diffusion supports a negative prompt field where you list what you do not want. Beginners who skip this consistently get extra limbs, blurry faces, watermarks, and low-quality textures. A basic negative prompt like 'blurry, low quality, extra fingers, watermark, deformed, ugly' removes a large class of common artifacts immediately.
Forgetting Composition and Framing
Without framing instructions, the model decides the crop and composition. Add terms like 'close-up portrait,' 'wide angle shot,' 'bird's eye view,' 'centered subject,' or 'rule of thirds' to direct the composition intentionally. This is especially important for portraits where you want a specific framing, such as a headshot versus a full-body view.
Using Natural Language Instead of Keywords
Stable Diffusion was not trained on conversational text. Writing 'Please create an image of a warrior that looks powerful' performs worse than 'powerful warrior, battle-worn armor, fierce expression, dynamic pose.' Strip out conversational filler and compress your intent into noun-adjective keyword clusters for better results.

Related queries

Frequently asked questions

How long should a Stable Diffusion prompt be for beginners?

A good beginner prompt is typically 20 to 50 words. Short enough to stay focused, long enough to give the model clear direction. Include your subject, setting, lighting, style, and 2-3 quality modifiers. Longer is not always better. Redundant words dilute the weight of important tokens.

What words make Stable Diffusion images look more realistic?

For photorealism, use terms like 'photorealistic,' 'shot on Canon EOS R5,' 'f/1.8 aperture,' 'natural lighting,' 'hyperdetailed,' and 'sharp focus.' Specifying a real camera model and lens is especially effective because the training data associates those terms with actual photographs rather than illustrations.

Do prompt examples work the same on SDXL and older SD 1.5 models?

Not exactly. SDXL understands more natural language and longer prompts better than SD 1.5. Prompts optimized for SD 1.5, which often rely on keyword-dense shorthand, may produce different results on SDXL. For SDXL, you can write more descriptively. For SD 1.5, tighter keyword clusters tend to work better.

What is a negative prompt and do beginners need one?

A negative prompt tells the model what to exclude from the image. Yes, beginners absolutely need one. At minimum use 'blurry, low quality, deformed, extra limbs, watermark, text' in your negative prompt. This alone eliminates a large percentage of the artifacts that make beginner outputs look amateurish.

Can I copy prompts from other people and expect the same results?

Not always. Output depends on the model checkpoint, sampler, CFG scale, seed, and resolution you use, not just the prompt text. A prompt that produces a stunning image on one setup may look completely different on another. Use shared prompts as a starting point and expect to adjust settings and wording for your specific environment.

What is the difference between a style prompt and a subject prompt?

The subject prompt describes what is in the image, such as 'a knight standing on a cliff.' The style prompt describes how it is rendered, such as 'oil painting, dramatic lighting, renaissance style.' Both are necessary. Beginners often write detailed subjects with no style guidance and get technically correct but visually flat results. Always specify both.