My prompt was refused

Safety filters, and how to rephrase.

Each model has its own safety filter. A prompt can be refused for violence, nudity, real people, brands or minors, sometimes because of a single ambiguous word.

• Remove names of real people and brands.
• Describe the scene rather than the sensitive element.
• Try another model: filters differ.

A moderation refusal refunds the generation debit. Edit the prompt before trying again.