I've been experimenting a lot with applications for local LLMs.

This one makes a ton of sense, and might even be native in Chrome at some point.

It's a Chrome Extension, that takes each new page automatically or manually and runs it through a local run LLM model.

We look for 6 different signals, to try to identify if its a potential phishing attack or not.

Local LLM because of privacy.

I've attached a demo video on the repo.