Negative prompt
A negative prompt is text that tells an image, video, or audio generator what to avoid producing — the opposite of the main prompt.
Used heavily in Stable Diffusion, Flux, Midjourney (via --no), Sora, and Kling, the negative prompt is a separate input that pushes the diffusion model's sampling away from certain features. Typical entries: "blurry, low quality, extra fingers, text, watermark, deformed hands". Negative prompting is most effective on open-weight diffusion models where the user controls sampling; closed APIs (Sora, Veo) often expose a constrained version. Good negative prompts are short and concrete — long lists of negatives cancel each other out and can degrade quality.
When to use negative prompt
- Removing common diffusion artifacts (extra limbs, garbled text, watermarks).
- Steering away from a style you don't want (e.g. "--no anime" on Midjourney).
Common mistakes
- Stacking 30 negatives — they fight each other and quality drops.
- Using negatives as a substitute for a precise positive prompt.
FAQ
What is negative prompt?
A negative prompt is text that tells an image, video, or audio generator what to avoid producing — the opposite of the main prompt.
When should I use negative prompt?
Removing common diffusion artifacts (extra limbs, garbled text, watermarks). Steering away from a style you don't want (e.g. "--no anime" on Midjourney).
What are the most common mistakes with negative prompt?
Stacking 30 negatives — they fight each other and quality drops. Using negatives as a substitute for a precise positive prompt.
Related terms
- Prompt engineering — Prompt engineering is the practice of designing input text that reliably steers a large language model toward a specific output.
- Diffusion model — A diffusion model is a generative neural network that creates images, video, or audio by iteratively denoising random noise toward a learned target distribution.
Last updated: 2026-06-01. Raw markdown: https://promtable.com/glossary/negative-prompt.md.