Nirantar was created to study continual learning in the context of building Automatic Speech Recognition (ASR) systems in practical scenarios where data collection happens incrementally.