Sightengine is a content moderation API spanning every format user-generated content actually arrives in: images with 120+ moderation classes, video and live streams, text, audio with profanity detection, and OCR plus QR code moderation.
The breadth is the point, because moderation gaps get found and exploited fast. A platform that screens images but not the text inside them, or the QR code in the corner of a photo, has left the obvious route open – and QR moderation in particular is a detail most vendors miss entirely despite it being a standard way to route users off-platform to something worse.
Live stream support matters for anyone whose risk is real-time rather than reviewable after the fact. Two things to plan for regardless of vendor: classifiers produce false positives and negatives at any threshold you pick, so build a human appeal path, and moderation thresholds carry cultural and legal context that no API infers for you – the policy decisions remain yours.
