A new speech recognition API has been added, which converts speech to text locally. It supports both real-time and batch transcriptions and processes input via microphone, as an audio stream, or from ...