Skip to main content
Interim results is a configuration option that enables “previews” of speech-to-text output to be returned over the web socket with lower latency than final transcript results.

Feature availability:

/transcribe

/stream

/transcripts

Additional cost may be incurred for use of this feature due to increased transcript output. If, however, you want to present users visual feedback of speech recognition in progress, then low latency interim results can be used to highlight active processing.

​
How it Works

During active dictation with configuration "interimResults" = true, transcripts will be returned with "isFinal" = false: 
{
"type": "transcript",
"data": {
    "text": "patient reports mild chest pain period",
    "rawTranscriptText": "patient reports mild chest pain period",
    "start": 0.0,
    "end": 3.2,
    "isFinal": false
  }
}
Interim results will have the same text listed in transcript responses for text and rawTranscriptText as post-processing for commands, punctuation, and formatting are skipped.

​
Tips for Use

  • Interim results can be used to present visual feedback to the user, during active dictation, to bring reassurance that audio is being processed.
  • This also supports users ability to monitor quality of dictation in real time, making it easier to spot potential errors or edits to make.
  • Be sure to only display interim results temporarily as a preview. Final results output should be accepted for use in the target workflow.
  • Due to the shorter context recognition window for interim result processing, some transcript previews may include inaccuracies that are corrected in the final results.
Additional cost may be incurred for use of this feature due to increased transcript output