Content moderation returns a verdict, scores, and time-stamped segments

The content moderation endpoints now return a richer result. A moderation result is a moderation object with an overall verdict, the modalities analyzed, the strongest confidence scores per category, and time-stamped segments that pinpoint each flagged moment.

What’s updated:

• Whole-media verdict and scores: read moderation.verdict and moderation.scores for a quick allow, review, or block decision.

• Time-stamped segments: each entry in moderation.segments carries a modality, a start and end time in seconds, a riskScore, and the categories that triggered it, so you can act on the exact moment instead of the whole asset.

• Video and audio coverage: modalities reports which tracks produced findings (video, audio, or both).

Use the video.media.ai.moderation.ready webhook to know when the result is ready.

API reference:

• Moderate video content and detect profanity: start a moderation job on a ready media asset.
• Get video moderation results: fetch the verdict, scores, and segments.

Guide: Content moderation