Media Moderation

Media Moderation is a separate, optional layer from text and link moderation — it classifies photos, and video where enabled, specifically for explicit/sexual content. It does not moderate media for other reasons (e.g. copyright, general appropriateness) — only this one category.

A high-confidence explicit result is enforced immediately (message deleted, escalating enforcement on repeats), the same severity as a high-confidence scam/phishing link. A borderline or uncertain result follows the same graduated warning path as unsolicited promotion.

Video classification is a separate capability from photo classification and is enabled per account where available — check your plan/account for current availability.

Media Moderation is billed separately from ordinary AI usage, in Media Moderation Units, with its own included allowance per plan — see "Plans and Billing."

This is distinct from text/link moderation, which covers spam, phishing/scam links, and unsolicited promotion — see "Moderation overview."