How do I use negative prompts in an image generation endpoint?

Negative prompts are usually used to describe what you don’t want in the generated image. In an image generation endpoint, you typically pass them as a separate field alongside the main prompt.

Basic idea

• Prompt: what you want
• Negative prompt: what to avoid

Example:

{
  "prompt": "a photorealistic portrait of a woman in a red dress",
  "negative_prompt": "blurry, low quality, extra fingers, distorted face"
}

Typical usage

Depending on the API, the parameter may be named:

• negative_prompt
• negativePrompt
• avoid
• exclude

Example request

response = client.images.generate(
    model="your-model",
    prompt="a futuristic city at sunset",
    negative_prompt="dark, foggy, low detail, blurry"
)

Tips

• Keep the negative prompt specific and concise.
• Common exclusions:
  • blurry
  • low quality
  • bad anatomy
  • extra limbs
  • watermark
  • text
• Too many negatives can sometimes hurt image quality or make the result less creative.

If your endpoint doesn’t support negative prompts

Some image generation APIs don’t offer a dedicated negative prompt field. In that case, you may need to:

• incorporate exclusions into the main prompt, or
• use model-specific guidance parameters if available.

If you want, I can show you how to do this for a specific API or SDK.