[necubi]

There's also http://hnhiring.me which I wrote a couple of years ago. It's a bit more flexible with filtering, but uncurated.

Incidentally, HN is up at the top of painful websites to scrape. All comments are at the same level of hierarchy, and the appearance of threading is accomplished by a spacer <img> with a width set according to the nesting depth. There are also no absolute times, and the times that are there slowly lose resolution. In order to get a comment time that's accurate to the minute, you have to load the page twice an hour and store the post id to time mapping.

[perspectivezoom]

Yep, the same level thing is a pickle. Like you, I'm determining top level comments via the same "s.img" indentation image with. Feels terrible, but it works.

Same with the comment time. Looked into the Sort by Recently Added request, and there's no good way to do it without polling.

Regex Filter added

[davedx]

Wish I had known about that before, spent so much time entering "site:news.ycombinator.com seeking freelancer" then trying to find the right month and year! Thanks for making that, will be using it in the future. :)