Mistral Moderation

Mistral Moderation is a content moderation model from Mistral AI designed to classify text content for safety and policy compliance. It detects categories including hate speech, violence, sexual content, self-harm, harassment, and other potentially harmful content, providing structured moderation decisions.

The model supports multiple languages and provides granular category-level scores, enabling applications to implement nuanced moderation policies rather than simple binary safe/unsafe decisions. It can be used for both input screening and output filtering in AI systems.

Mistral Moderation is essential for production AI deployments that need to ensure content safety, regulatory compliance, and brand protection in user-facing applications.