@aws-sdk/client-transcribe-streaming

Real-time Transcription

This feature allows you to start a real-time transcription session. The code sample demonstrates how to set up a TranscribeStreamingClient, create an audio stream, and start the transcription process. The transcription results are logged to the console as they are received.

const { TranscribeStreamingClient, StartStreamTranscriptionCommand } = require('@aws-sdk/client-transcribe-streaming');
const { Readable } = require('stream');

const client = new TranscribeStreamingClient({ region: 'us-west-2' });

const audioStream = new Readable();
audioStream._read = () => {}; // No-op

const command = new StartStreamTranscriptionCommand({
  LanguageCode: 'en-US',
  MediaSampleRateHertz: 16000,
  MediaEncoding: 'pcm',
  AudioStream: audioStream
});

const response = await client.send(command);

response.TranscriptResultStream.on('data', (event) => {
  console.log(event.TranscriptEvent.Transcript.Results);
});

audioStream.push(audioBuffer); // Push audio data to the stream

Handling Transcription Events

This feature demonstrates how to handle transcription events. The code sample shows how to process partial and final transcription results from the TranscriptResultStream.

response.TranscriptResultStream.on('data', (event) => {
  event.TranscriptEvent.Transcript.Results.forEach(result => {
    if (result.IsPartial) {
      console.log('Partial transcript:', result.Alternatives[0].Transcript);
    } else {
      console.log('Final transcript:', result.Alternatives[0].Transcript);
    }
  });
});

Stopping the Transcription

This feature shows how to stop the transcription process by signaling the end of the audio stream. The code sample demonstrates how to push a null value to the audio stream to indicate that no more audio data will be sent.

audioStream.push(null); // Signal the end of the audio stream