I have a problem with all these website-blocking tools that I want them to behave differently than they do, perhaps you can suggest something?

The issue is that I want to block only the main, 'browsing' pages while allowing deep links to that same site. For example, I want to block reddit, when googling some question turns out an answer that's on a niche subreddit I've never seen, I need to be able to read it. However, the site blockers I've tried require me to unblock the whole site or turn them off, and thus become pointless.

If you have the patience to manually blacklist URLs, you can do this pretty easily with a regex-based site blocker. (My favorite for Chrome is Simple Blocker, which doesn't document that it uses regexes but in fact does: https://chrome.google.com/webstore/detail/simple-blocker/akf...). For example, the following blacklist regex patterns block the Reddit front page and the main page of any subreddits and multireddits, but still let you navigate to individual threads:

    https://(www|np).reddit.com/?$
    https://(www|np).reddit.com/r/[^/]+/?$
    https://(www|np).reddit.com/user/[^/]+/m/[^/]+/?$

Thank you, this has been really helpful for me and my habits.

Hmm... That is an interesting problem, but I've not found anything. You would need some sort of proof-of-work for it to make behavioral sense though.

I don't think that proof-of-work is needed - these are different models of operation; reading an article (which is ok) vs browsing for more articles (which is not).

With StayFocusd at least, you can whitelist subreddits while blocking the main site.

I want to blacklist only the main site - Stayfocusd was a particular example, I got some google results on subreddits that I never knew existed and couldn't have been in a whitelist.

I'm rather going for the reading vs scanning distinction - even there I would want the particular subreddit main page to remain blocked to prevent the procrastinating pattern of checking an interesting place to see if there's something new interesting. Similarly, the main news.ycombinator.com should be blocked but visiting particular pages from e.g. a weekly digest email would be useful.

Or have cron jobs that swap out /etc/hosts with pre-prepared replacements