azure-ai-contentsafety-py

You can detect harmful content in text and images by running multi-severity classification scripts that analyze input data for hate, sexual, violent, and self-harm material. This enables automated content moderation and filtering for online platforms.

Multi-severity classification for content moderation is the process of categorizing detected harmful text and images into different risk levels. It allows applications to apply varied filtering rules based on the severity of hate, violence, or self-harm content found.

Yes, you can create and manage custom blocklists to filter specific text content. This allows you to block predefined terms or phrases in addition to the automated harmful content detection provided by the Azure AI Content Safety SDK.

Yes, Azure AI Content Safety supports image classification for violent, self-harm, hate, and sexual content. You can analyze images using image classification to detect and filter harmful visual material on social media platforms or forums.

To analyze text for hate and violent content, you need the azure-ai-contentsafety Python dependency installed in your environment. You then configure your application to pass text inputs to the analysis scripts for multi-severity classification.