@tg When reading about Current a question popped up in my head around the categories I have to define per source: For some sources this might work, but others publish mixed content, sometimes breaking stories, sometimes articles. As iOS has built in offline/local natural language processing and machine learning support, could Current detect this category on its own? That way, you could mark a source as mixed and let Current find out on its own. I am not suggesting using cloud LLMs.