Separate positive and negative
Use the positive prompt for what you want, and reserve the negative prompt for artifacts and failure modes.
Free AI image prompt tool
Stable Diffusion Prompt Generator
Build Stable Diffusion and SDXL prompts for portraits, product shots, concept art, interiors, anime, and posters. Add your idea, choose model settings, then copy a positive prompt, negative prompt, and recommended generation settings.
Use case
Positive, negative, and settings prompts
Best for
SDXL, portraits, products, concept art, anime
Output
Positive prompt, negative prompt, settings
Quick starts
Aspect ratio
Advanced settings
Generated diffusion brief
professional studio portrait of a creative director wearing a black turtleneck, soft key light, subtle background gradient
SDXL with soft key light with subtle rim light
Stable Diffusion settings
Size
1024x1024
Model
SDXL
Sampler
DPM++ 2M Karras
Steps
35 steps
Positive prompt
A professional studio portrait of a creative director wearing a black turtleneck, soft key light, subtle background gradient, photorealistic, highly detailed, sharp focus, clean composition, 85mm lens, shallow depth of field, soft key light with subtle rim light, Keep the subject clear, preserve realistic proportions, and avoid cluttered backgrounds.
Negative prompt
low quality, low resolution, blurry, out of focus, bad anatomy, bad hands, extra fingers, missing fingers, distorted face, deformed body, duplicate limbs, long neck, cross-eye, poorly drawn eyes, text artifacts, watermark, logo artifacts, jpeg artifacts, noisy background, oversaturated, plastic skin, over-sharpened, messy composition
Recommended settings
Model: SDXL Size: 1024x1024 (1:1) Sampler: DPM++ 2M Karras Steps: 35 CFG scale: 7 Seed: random or lock after a good first result Hires fix: optional for final upscales
Prompt formula
How to make Stable Diffusion prompts more controllable
Stable Diffusion workflows are easier to tune when the positive prompt, negative prompt, model, sampler, steps, CFG scale, and image size are kept separate. This generator gives you a clean starting point for each piece.
Match model and style
Choose realistic wording for photoreal models, anime wording for anime models, and broader scene language for SDXL.
Start with stable settings
Sampler, steps, CFG, and size matter. Start simple, then lock the seed after a promising result.
Examples
Stable Diffusion prompt examples
Portrait Prompt
A professional studio portrait of a creative director wearing a black turtleneck, soft key light, subtle background gradient, photorealistic, highly detailed, sharp focus, 85mm lens, shallow depth of field
Product Prompt
A premium skincare bottle on a white ceramic surface with water droplets, soft reflections, clean studio lighting, premium product photography, studio lighting, polished surfaces, commercial finish
Concept Art Prompt
A futuristic city gate with floating transit lanes, giant glass towers, mist, and tiny figures for scale, cinematic concept art, atmospheric depth, cinematic lighting, detailed environment
SEO guide
Best practices for Stable Diffusion prompts
Keep positive prompts descriptive but organized. Start with the subject, then add style, lighting, camera, quality, and details. Avoid mixing too many unrelated aesthetics in one prompt.
Negative prompts should target common failure modes. Bad anatomy, extra fingers, blurry details, watermarks, text artifacts, and noisy backgrounds are useful defaults for many Stable Diffusion workflows.
What is a Stable Diffusion prompt generator?
A Stable Diffusion prompt generator creates a positive prompt, negative prompt, and recommended settings for image generation workflows such as SDXL, Stable Diffusion 1.5, realistic models, and anime models.
How do I write better Stable Diffusion prompts?
Describe the subject first, then add style, quality terms, camera or composition, lighting, and any extra constraints. Keep the negative prompt focused on common artifacts you want to avoid.
What should go in a negative prompt?
Use negative prompts to reduce blurry images, bad anatomy, extra fingers, distorted faces, duplicate limbs, watermarks, text artifacts, noisy backgrounds, and messy composition.
What settings should I start with?
A practical starting point is SDXL, DPM++ 2M Karras, 25-35 steps, CFG around 7, and a size that matches the intended aspect ratio.