AI Feed Shield per Prashant Prabhakar

Built as a next-generation digital wellbeing tool, AI Feed Shield puts you back in control of your social media experience. Instead of relying on slow, platform-wide moderation that often misses the mark, this extension acts as your own personal AI bodyguard, scanning and filtering out harmful content in milliseconds before it ever reaches your screen.

Key Features:

Real-Time Toxicity Engine: Powered by a custom fine-tuned BERTweet AI model, the extension instantly detects and blocks hate speech, harassment, and abusive language in posts and comments.
Hybrid Image Safety: Combines a state-of-the-art Vision Transformer (ViT) with intelligent semantic heuristics to accurately detect and blur NSFW, explicit, and graphic imagery.
Native UI Integration: When harmful content is detected, it is seamlessly replaced with a sleek, platform-native "System Warning" (matching X/Twitter's dark mode), keeping your feed looking clean and professional.
Granular Control: An interactive dashboard allows you to toggle the Master Shield, Image Filtering, and Text Filtering on or off at any time.
Privacy First: Your feed data is processed securely via a local tunnel architecture, meaning your personal browsing history is never permanently collected or stored on third-party marketing servers.
Take control of your mental health and digital wellbeing with the ultimate AI-powered content filter.