Comment by econ

I had a good few feeds that had errors that prevented them from parsing. I examine the flaws and parse them anyway. Then one day I discover a website had dropped rss support long ago but I had it in my aggregator?? I open the feed url I was using and it just redirected to the index.html took a minute to realize what was going on. If it couldn't find <item> or anything like it it would look for anything similar, if it couldn't find <link> or <guide> it would search for <a>, if it couldn't find a <title> it would take the text from the <a> or use the url and lastly if it couldn't find or parse <pubdate> it would look inf the item url had something like /2025/ in it, prerably /months/ and /day/ with it.

So that was what was going on. It could find links on the frontpage and it could parse titles and dates from those.

You apparently don't need feeds. No AI required.