Looking for russian content on russian websites, will be very similar to english content on english websites. Unfortunately mixing the languages is a whole other topic. It's a hard problem, as the concept of content matching begins to look more like machine learning (sentence level fuzzy pattern matching, word2vec, etc.), rather than simple pattern matching, as one cannot simply "undo" a translation into the exact content it started as. You can try this yourself by going in google translate, and translating sentences back and forth even in the same language -- the meaning gets garbled.
Even if you could undo a translation, one would have to identify the content language, and then for all languages, search for the content in the translation. Even if we only add one language, this doubles the price of running cheetah (assuming API translation services are free -- they are not).
So, it's a hard problem, and one that I don't really have a way to automate. :(
http://minds.com has an open source translator.
I don't know how efficient it is.
That's probably the way to start, with an offline translation program. However, it unfortunately still increases the search space and cost. :(
I understand. Too bad :(