title: "API Mistral Moderation | Mistral AI"

Mistral AI has released a new content moderation API to provide system-level safety guardrails for downstream applications. This tool is an LLM-based classifier trained to categorize text inputs into nine distinct policy categories, addressing harms like unqualified advice and PII. The API is natively multilingual and offers two endpoints, one for raw text and another for classifying messages within a conversational context. The company shares performance benchmarks for the classifier, which is the same one used to power moderation in its Le Chat service.