Random Reaper extracts links from dmoz.org smartly, and discards all junk links. Links are then stored in a MySQL database, so if you run a web server it will be faster than dynamically fetching links and modifying them. The difference between Random Reaper and its competitors is that RR uses a new technique in getting links where it can differentiate between rogue links and links that form the main structure of the site. You don't need to enter a search phrase, all you do is just select a category, subcategory and get a random links from there.

There are some intentional incompletes in the code so dmoz.org doesn't get flooded by RR if lots of people use it. If you really want to use RR to get every single category and link from dmoz.org then just fill in the blanks. RR is open source distributed with GPL.