Generate Prompt

Negative Prompts for AI Images

Copy negative prompts for AI images to reduce blurry results, bad hands, text errors, watermarks, distorted faces and low-quality outputs.

Negative Prompt Problem Solver Table

Use this table when you already know what went wrong in the image. Copy the matching negative prompt, then keep only the phrases that fit your prompt instead of pasting every row at once.

ProblemNegative promptBest for
Bad handsbad hands, extra fingers, missing fingers, fused fingers, distorted palms, broken knuckles, unnatural wristsportraits, selfies, fashion shots
Text errorsunreadable text, random letters, misspelled words, fake typography, broken headline, text fragments, watermarkposters, thumbnails, packaging
Product shapewarped product shape, bent packaging, crooked label, duplicate product, melted edges, messy reflectionsproduct photos, ecommerce images
Distorted facesdistorted face, asymmetrical eyes, warped mouth, duplicated face, uncanny smile, plastic skin, over-smoothed skinrealistic portraits, avatars
Messy backgroundscluttered background, distracting objects, random props, background artifacts, busy composition, low detail backdropportraits, product photos
AI logosmisspelled text, unreadable lettering, warped symbol, extra icon, fake watermark, jagged edges, noisy backgroundlogo concepts, brand marks
Postersunreadable headline, random letters, misspelled title, broken typography, cropped title, watermark, low resolutionmovie posters, event posters
Midjourney style control--no text, watermark, signature, extra fingers, distorted face, blurry, low quality, duplicate subjectstylized art, cinematic images
Flux realismplastic skin, waxy face, distorted eyes, bad hands, overprocessed detail, unnatural texture, blurry edgesrealistic portraits, editorial photos

Best Negative Prompts by Use Case

These long-tail negative prompts are written for common image generation jobs: portraits, products, logos, posters and model-specific workflows.

Best negative prompt for realistic portraits

Use this when the goal is a natural headshot, selfie, avatar or editorial portrait. It focuses on faces, hands, skin texture and image quality without removing useful detail.

Best negative prompt for realistic portraits
bad hands, extra fingers, missing fingers, fused fingers, distorted palms, distorted face, asymmetrical eyes, warped mouth, duplicated face, plastic skin, waxy skin, over-smoothed skin, blurry facial features, low resolution, noisy, oversharpened, unnatural lighting

Best negative prompt for product photos

Best for ecommerce images, clean product shots, packaging mockups and catalog visuals where shape, edges and label readability matter.

Best negative prompt for product photos
warped product shape, bent packaging, crooked label, duplicate product, melted edges, messy reflection, cluttered background, random props, unreadable label, fake logo, low quality product edges, blurry product, distorted perspective

Best negative prompt for AI logos

Use for logo concepts and simple brand marks. AI logo text still needs manual checking, so keep the negative prompt focused on lettering, geometry and clutter.

Best negative prompt for AI logos
misspelled text, unreadable lettering, random letters, fake brand name, warped symbol, extra icon, broken geometry, jagged edges, watermark, signature, busy background, low resolution, blurry logo

Best negative prompt for posters

Useful for movie posters, event posters, music covers, social ads and thumbnails where typography and composition often break.

Best negative prompt for posters
unreadable headline, random letters, misspelled title, broken typography, watermark, signature, fake credits, cropped title, cluttered layout, low resolution, jpeg artifacts, blurry focal point, distorted face

Best negative prompt for ChatGPT image generation

ChatGPT image prompts may not always use a separate negative field, so write the constraints as plain English at the end of the prompt.

Best negative prompt for ChatGPT image generation
Avoid blurry hands, extra fingers, distorted faces, unreadable text, visible watermark, fake signatures, warped product shapes, low-resolution artifacts, messy backgrounds and duplicate subjects.

Best negative prompt for Midjourney

Midjourney commonly uses the --no parameter. Keep it short and tied to the failure you actually see, especially for stylized portraits and cinematic scenes.

Best negative prompt for Midjourney
--no text, watermark, signature, extra fingers, bad hands, distorted face, duplicate subject, blurry, low quality, jpeg artifacts, messy background

Best negative prompt for Flux

Flux is often used for realistic detail, so this negative prompt protects skin, eyes, hands, edges and natural texture.

Best negative prompt for Flux
plastic skin, waxy face, distorted eyes, asymmetrical face, bad hands, extra fingers, unnatural texture, overprocessed detail, blurry edges, low resolution, watermark, unreadable text, warped objects

What Is a Negative Prompt?

A negative prompt tells an image model what to avoid. It is most useful when the model repeatedly creates the same problem: bad hands, extra fingers, distorted faces, unreadable text, watermarks, blurry image quality, low resolution, warped product shape, messy background or broken anatomy. Keep the negative prompt specific and connected to the image you are making.

Negative Prompts for Bad Hands

Hands fail often because fingers overlap and small shapes are easy to distort.

Negative Prompts for Bad Hands #1
bad hands, extra fingers, missing fingers, fused fingers, broken knuckles, distorted palms, unnatural wrist, blurry hands, low detail hands

Negative Prompts for Faces

Use these for portraits, profile images and realistic people.

Negative Prompts for Faces #1
distorted face, asymmetrical eyes, warped mouth, duplicated face, uncanny expression, plastic skin, over-smoothed skin, blurry facial features

Negative Prompts for Text and Watermarks

Useful for posters, packaging mockups, thumbnails and social posts.

Negative Prompts for Text and Watermarks #1
unreadable text, random letters, watermark, signature, logo artifacts, fake brand text, misspelled label, text fragments

Negative Prompts for Product Photography

Protect product shape, label clarity, reflections and clean backgrounds.

Negative Prompts for Product Photography #1
warped product shape, bent packaging, crooked label, duplicate product, messy reflection, cluttered background, low quality product edges

Negative Prompts for Realistic Portraits

Combine face, skin, hands and quality constraints for portraits.

Negative Prompts for Realistic Portraits #1
distorted face, asymmetrical eyes, warped mouth, duplicated face, uncanny expression, plastic skin, over-smoothed skin, blurry facial features, bad hands, extra fingers, missing fingers, fused fingers, broken knuckles, distorted palms, unnatural wrist, blurry hands, low detail hands, blurry, low resolution, noisy, oversharpened, compression artifacts, bad lighting, muddy colors, washed out highlights

Negative Prompts for ChatGPT Images

Write constraints as plain English if there is no separate negative field.

Negative Prompts for ChatGPT Images #1
ChatGPT image constraint: avoid blurry hands, distorted faces, unreadable text, visible watermark, warped products and low-resolution artifacts

Negative Prompts for Stable Diffusion

Stable Diffusion workflows often support a dedicated negative prompt field.

Negative Prompts for Stable Diffusion #1
Stable Diffusion negative: lowres, blurry, bad anatomy, bad hands, extra fingers, watermark, text, jpeg artifacts, worst quality, low quality

How to Combine Positive and Negative Prompts

Write the positive prompt first: subject, scene, style, lighting, camera and desired details. Then add only the negative constraints that match the likely failure points. A product photo may need warped label and messy reflection negatives, while a portrait may need bad hands, distorted eyes and plastic skin negatives. Avoid adding every possible negative word to every prompt because broad negatives can remove useful detail.

Related Prompt Categories

FAQ

Do negative prompts always improve images?

No. They help most when they target predictable errors, but too many broad negatives can make the image flat.

Can I use negative prompts with ChatGPT images?

Yes. If there is no separate negative field, write the constraints directly in the prompt.

What is the best negative prompt for hands?

Use focused terms such as bad hands, extra fingers, fused fingers, missing fingers and distorted palms.

Should product prompts use negatives?

Yes, especially for warped packaging, crooked labels, messy reflections and duplicate products.

Are Stable Diffusion negatives universal?

No. Treat CFG, steps, sampler and negatives as workflow concepts that need testing with your model.

What negative prompts are

A negative prompt tells the image model what to avoid. It is useful when the first result has predictable errors: extra fingers, warped faces, fake text, distorted product packaging, messy backgrounds, watermarks or blurry details. Use it as a cleanup tool, not as a replacement for a clear main prompt.

When not to overuse negatives

Long negative prompts can fight the main prompt and remove useful detail. Start with the few problems you actually see, then add more only when the same failure repeats. For ChatGPT image generation, write constraints as plain English when there is no separate negative prompt field.

Examples by category

  • Portraits: bad hands, extra fingers, distorted eyes, plastic skin, uncanny expression.
  • Product photos: warped packaging, crooked label, fake logo, messy reflection, duplicate product.
  • Posters and thumbnails: unreadable headline, random letters, fake signature, cluttered layout.
  • Realistic scenes: unnatural shadows, oversharpened texture, low-resolution artifacts, distorted background.

Explore related tools