The moderation API is powered by a fine-tuned version of Mistral AI's Ministral 8B and is designed to detect potentially harmful content across nine different categories, including sexual content ...