AI-Driven Toxic Comments Classifier

Project Overview

A custom Natural Language Processing project designed for social platforms, forums, and blogs. It scans user comments in real-time, categorizes them under threats, obscenity, insults, or identity hate, and flags them for admin approval before publishing.

Core Features & Functions

• API endpoint for real-time comment screening
• Multi-label classification ML models (LSTM/BERT)
• Interactive analytics panel displaying toxicity trends
• Automated email warnings for repeat offenders
• Custom word blocklist and moderation logs