A fully managed automatic speech recognition (ASR) service that converts speech into text using deep learning. It supports both batch and streaming transcriptions with features like punctuation, diarization, and language identification.