Word-level time accuracy, gpu accelerated transcribe & diarize, automatic speech recognition w/ whisper (large-v3).
