When speaking with Pi, if a single transcription request fails (e.g. timeout) the merged transcript does not automatically get submitted.
Expected Result:
When the user stops speaking, the transcript should be submitted. Any failed transcriptions within that should be omitted or replaced with '...'.
Actual Result:
Say, Pi continues to wait indefinitely for additional user input. Only after the user speaks again, and that transcription is received, is the final transcript merged and submitted.
When speaking with Pi, if a single transcription request fails (e.g. timeout) the merged transcript does not automatically get submitted.
Expected Result: When the user stops speaking, the transcript should be submitted. Any failed transcriptions within that should be omitted or replaced with '...'.
Actual Result: Say, Pi continues to wait indefinitely for additional user input. Only after the user speaks again, and that transcription is received, is the final transcript merged and submitted.