Skip to main content
Every moderation job evaluates content against three built-in categories. Each category is a boolean — true if detected, false if not.

​
Categories

​
NSFW

Sexually explicit or not-safe-for-work content detected. Includes nudity, sexual acts, and other adult material.

​
AI Adult General

Catch-all classifier for adult content not covered by the other built-in categories.

​
Underage

Content involving minors detected. The job is rejected like any other category violation. Repeated detections may be reviewed manually and can result in account termination at our discretion — see account termination.

​
Category output

Categories appear in the top-level categories object of a completed job. Per-client custom categories are nested under categories.custom:
categories.custom is only included when at least one custom category ran for the job. If you haven’t configured any custom categories, the custom key is omitted entirely.

​
How categories affect decisions

If any category (built-in or custom) is true, is_prohibited is true. All categories must be false for is_prohibited: false. See decision logic for details. For video moderation jobs, categories are merged across frames — if any single frame triggers a category, the entire video is flagged.