Moderation API is a hosted content moderation service covering text, images, video, and audio. You write your own guidelines in plain language, and its models apply those rules to incoming content with an explanation and a confidence score, across more than 120 languages through a single endpoint.
Alongside detection, the platform covers the operational side of trust and safety. Clear violations are handled automatically while a review queue collects the edge cases for a team, user records carry history and behavior patterns rather than single pieces of content, and if-then rules combine content scores with user signals to decide an action. Decisions a review team makes are fed back to adjust the models.
Features
- Guideline-driven detection: you describe what to moderate the way you would brief a new hire, and the models apply those rules with an explanation and a confidence score
- Multiple content types: text, image, video, and audio moderation through one platform
- Pre-trained models: more than 20 detection models are available without training your own
- Custom rules: if-then logic combines user signals, content scores, and trust levels into actions such as block, escalate, or auto-approve
- Conversation context: history is considered with each decision, so the same sentence is not scored identically in every situation
- Language coverage: more than 120 languages behind a single API
- Review queue: a team observes automated decisions and handles the cases automation leaves open
- User management: per-user records show first seen, last active, category scores, and actions such as cancel, suspend, or ban
- Adaptive learning: each review a team makes feeds back into the models without prompt changes or retraining