Description
Speech recognition workflows can use local or hosted engines through one Python interface. Developers can capture audio, process files, and send speech to supported recognition services depending on the backend. Microphone input and remote APIs may expose private speech content.