How to Use Negative Prompts

A negative prompt is a short list of things you do not want in your image. Used well, it cleans up common flaws, but used too heavily it can strip out detail you actually wanted.

Quick answer

A negative prompt tells the model what to leave out, such as extra fingers, blur, stray text, or clutter. You write it in a separate field or section, and the model treats those items as things to avoid. Keep the list short and specific so you remove flaws without flattening the whole image.

What a negative prompt does

A normal prompt pulls the image toward what you describe. A negative prompt pushes it away from what you list. Together they shape the result from both sides.

The model does not erase items in a literal sense. Instead, it lowers the chance those features appear. So a negative prompt is a nudge, not a hard rule, and it works best on common problems the model already knows how to avoid. If you are still shaping the main description, our notes on AI image prompt structure cover the positive side first.

What to exclude

Good negative prompts target known failure modes rather than vague worries. The table below lists common items and when each one helps.

You rarely need every row. Pick the two or three that match the flaw you keep seeing. To go deeper on the main description while you tune this list, our walkthrough on how to write AI image prompts pairs well with this page.

How to avoid over restricting

The biggest mistake is a giant negative list copied from somewhere else. Long lists can remove detail, dull colors, or make images look flat and lifeless. They also slow down your testing, because you cannot tell which item fixed the problem.

Start with an empty negative prompt. Generate an image, look at what went wrong, then add one item to fix it. Generate again. Add another item only if a new flaw shows up. This one-at-a-time habit keeps your list lean and your results rich. If a small fix is all you need, light cleanup with our image enhancement guide can sometimes solve the issue without any negative prompt at all.

Checklist

• Start with no negative prompt and add items only as needed.
• Target real flaws you see, not flaws you fear.
• Use plain, specific words like "blurry" or "extra fingers."
• Add one item at a time so you know what each one does.
• Keep the list short, usually under six items.
• Remove any item that does not change the result.
• Watch for flat or dull images, which can mean the list is too long.

Example

Here is a positive prompt followed by a short, focused negative prompt for a portrait. The negative list targets only the flaws portraits tend to show.

Positive: A friendly portrait of a woman smiling, soft studio light,
clean gray background, sharp focus, natural skin tone

Negative: extra fingers, deformed hands, blurry, text, watermark

For a product shot, you would swap in product flaws instead, such as "reflection glare, dust, busy background." Match the negative list to the kind of image you are making, and try our AI image generator to test how each change shifts the output.

FAQ

Do I always need a negative prompt?

No. Many images come out clean without one. Reach for a negative prompt only when a specific flaw keeps appearing.

Can a negative prompt remove anything?

Not reliably. It lowers the chance of common, well-known features showing up, but it cannot force-delete every unwanted item. For stubborn problems, change the positive prompt too.

Why did my image get worse after a long negative list?

Long lists can strip out useful detail and make images look flat. Cut the list down to the few items that fix real flaws and test again.

Where do I put the negative prompt?

Most tools have a separate negative prompt field. If yours does not, check whether it supports a marker in the main prompt, and keep the negative items clearly separated from the positive ones.