This paper introduces the motivations and first results of the creation of Czech RapCor project, mainly the constitution process of Czech RapCor Boosted v1 (Czech RCB), a specialized corpus of Czech rap lyrics designed for sociolinguistic and NLP research. The corpus highlights distinctive linguistic features, such as written colloquialism, frequent use of vulgarisms, and non-standard forms, which pose challenges for traditional NLP tools. Preliminary results demonstrate the corpus’s potential for studying authentic spoken language in written form, offering insights into rap culture and sociolinguistic phenomena.