Corpus of Spoken Hindi (COSH)

COSH (Corpus of Spoken Hindi) is a large-scale web corpus consisting of approximately 200 million words collected from internet texts written in Devanagari script. The corpus is designed for linguistic research on Hindi, including word frequency analysis, collocations, grammatical patterns, and usage variation. A dedicated concordancer, COSH Conc, enables users to explore the corpus efficiently. COSH is currently used by researchers and students, especially in India, in the fields of corpus linguistics and computational linguistics. The COSH concordancer was developed with technical support from external developers.