Tried this for awhile with my girlfriend who speaks Mandarin as her first language. It does seem to be slightly too picky, often calling my pronunciation third tongue when it should be one of the others, despite her saying it was perfectly fine to her ear. But on the other hand it called her pronunciation perfect all but one time, which seems pretty encouraging.
I'm a guy, and it called all of my tones "3rd tone" when said in my normal voice, but when I spoke in a cartoonishly high falsetto, it gave me perfect marks.