MaSCoT is a tool developed to facilitate searching, extracting and analyzing information contained in large, richly-annotated speech corpora developed in Praat.
MaSCoT takes a TextGrid and a LongSound object as input, and allows users to perform sophisticated searches on these objects using regular expressions and multiple-tier conditions. MaSCoT's output always includes a report detailing the search results and the search and extraction options used. Optionally, it can also extract and label TextGrid fragments (to individual TextGrid files) and/or audio (to individual WAV files) from any tier related to the search results.
This version processes all TextGrid/WAV file pairs in a user-specified directory recursively, to a depth of 1 sub-directory.
More information here