The purpose of this paper is to segment carnatic music recordings into individual items for archival purposes using applauses. A concert in carnatic music is replete with applauses. These applauses may be inter-item or intra-item applauses. A property of an item in carnatic music, is that within every item, a small portion of the audio corresponds to the rendering of a composition which is rendered by the entire ensemble of lead performer and accompanying instruments. A concert is divided into segments using applauses and the location of the ensemble in every item is first obtained using Cent Filterbank Cepstral Coefficients (CFCC) combined with Gaussian Mixture Models (GMMs). Since constituent parts of an item are rendered in a single raga, raga information is used to merge adjacent segments belonging to the same item. Inter-item applauses are used to locate the end of an item in a concert. The results are evaluated for fifty live recordings with 990 applauses in total. The classification accuracy for inter and intra item applauses is 93%. Given a song list and the audio, the song list is mapped to the segmented audio of items, which are then stored in the database. © 2013 International Society for Music Information Retrieval.