Guides

Negative Prompts Explained

Negative prompts tell an AI image tool what to leave out. How they work in Stable Diffusion and Midjourney, with examples.

A negative prompt lists what you do not want in the image. It helps remove common problems such as blur, clutter or unwanted objects.

Where they work

Some tools have a separate negative prompt box. Stable Diffusion apps usually do. In Midjourney, you add --no followed by what to exclude. Many other generators have no negative option, so describe what you do want instead.

Examples

For Stable Diffusion apps, put this in the negative prompt field:

blurry, low quality, extra fingers, distorted hands, watermark, text

In Midjourney, add it to the end of your prompt:

a minimalist living room, natural light, wooden floor --no clutter, people

Tips

  1. Keep the list short and relevant; long lists can cause odd results.
  2. Describe what you want first. Negative prompts are for cleaning up.
  3. Change one thing at a time to see what each term does.
  4. If a tool ignores negatives, rewrite the main prompt positively, for example “empty room” instead of “no people”.

Learn the basics in How to write better AI image prompts and see Midjourney parameters explained.

stay in the loop

New guides, prompts and AI model news by email