Moderate Adult Content Using Azure AI Computer Vision

Quick Answer

Detecting adult or racy content is one of several distinct visual features the Analyze Image API can be asked to evaluate, and it has to be explicitly requested — passing the 'adult' parameter when calling the API tells Computer Vision to run this specific classification alongside whatever other analysis is requested, such as tagging or object detection. The API returns both a boolean flag indicating whether the content was classified as adult or racy and a confidence score for that classification, giving a moderation system enough information to decide not just whether to act but how confidently to act, which matters for setting an appropriate threshold rather than treating every borderline case identically. This is Computer Vision's answer to content moderation for still images specifically, distinct from Content Safety, which handles a broader range of harm categories including text, or Video Indexer's people detection, which tracks individuals rather than classifying image content. Any scenario asking specifically to flag adult or racy imagery, as opposed to a broader set of harmful content categories or a different media type, is pointing at the Analyze Image API's adult content parameter.