BSC-CNS works on the development of Speech Technologies to count with voice recognisers, encoders and text-to-speech models in order to make easier the adoption of those technologies.

The speech technologies are developed following accurately the ethical and legal framework, and the training process is based on data protection and privacy principles. BSC-CNS has developed text-to-speech models such as Matxa-TTS which integrates the Catalan dialect diversity, and new speech recognition models based on Whisper architecture that cover different needs of the community.