OP of Unim.press here-- yep, extracting text is pretty hard, though I think Instapaper and Pocket have good tech in that space. Certainly more than a weekend project's worth of work, which is why most of my focus personally was on the layout / visual fidelity. But it's cool to see that you put more time into extracting the content. I think for HN which is mostly longform online writing it definitely adds value.