PhobiaBlocker

PhobiaBlocker is a Chrome extension built for people who want to browse the web without unexpectedly running into images that trigger a phobia. You pick the objects you want hidden from a long list of real-world categories derived from ImageNet classes, and the extension uses machine learning models running locally on your machine to detect and block matching images as pages load. Because the analysis happens on your own computer, no browsing data is sent to an external endpoint.

Beyond automatic blocking, PhobiaBlocker gives you fine-grained manual controls. You can block or unblock a specific image from the right-click menu, block every image on a site, or write regex rules that automatically trust or block images based on their src attribute. Trusted website lists and a local decision cache help keep browsing fast, and a quick option lets you block spider-like and insect-like images in one click.

The extension supports a speed-versus-accuracy tradeoff, risk management settings, and both dark and light themes. It is designed around privacy by default: images are analysed locally, caches can be disabled, and the developer states that no data is collected or sold to third parties.