zym9863/comment-sentiment-filter

This project is a BERT-based sentiment analysis and content filtering service designed to analyze Chinese and English comments, categorizing them as toxic, negative, neutral, positive, or constructive. It's built for developers and content moderators who need to automatically filter and analyze user-generated comments, featuring both API endpoints and a frontend interface for easy interaction. The service uses a multilingual BERT model and supports both single and batch comment analysis with configurable thresholds for toxicity and constructiveness detection.