We take content safety seriously. Every output goes through multiple layers of protection:
What we block: Hate speech, harassment, explicit content, self-harm content, weapons instructions, medical/legal/financial advice, doxxing, political persuasion, and defamation.
Satire rules: All satirical content is labeled SATIRE / FICTION. We never mention real people, politicians, celebrities, or organizations. We never use realistic breaking-news styling.
Safety layers: System-level AI safety policy, output keyword filtering, named entity detection, automatic regeneration with stricter constraints, and deterministic safe fallbacks.
Reporting: Every output has a Report button. Reported content is logged for review. No personal information is stored.
Rate limits: To prevent abuse, usage is limited per session.
If you encounter anything that doesn't meet these standards, please use the Report button. We appreciate your help keeping this space fun and safe for everyone.