Church Slavonic corpus, opened in 2012, can dramatically improve the accuracy and objectiv¬ity of Church Slavonic language studies. The corpus differs from a collection of texts by special markup (grammatical, structural, metatextual) and search facilities. Existing grammars and dictionaries are of historical or regulatory nature and often do not match the facts reflected in real texts. This paper discusses the problem of spelling representation and grammatical tagging of texts in the corpus and its usage for studying grammatical variation.