Partial/interim transcription result. These are emitted during real-time transcription as the user speaks and may change as more audio is processed.
Confidence score from 0.0 to 1.0, if available
The current transcribed text (may change in subsequent updates)