www.contestado.com.br and www.gigabusca.com.br are live mirrors, but www.tiosam.com and www.ebaita.com doesn't seem to be. Data is from January. They state "Articles are cached copies extracted from Wikipedia, the Free Encyclopedia". Question is how do they cache it... :P Their use of ~ instead of : (and breaking back link to wikipedia for namespaces != 0) would suggest static dumps, but they include pages that they don't include. As theit source include the parser cache key, i think they crawled. Gives credit to wikipedia and informs about GFDL :-) It also asks to mention the web on future references ( they add nothing to the wikipedia content! :-( )