Am I using Text to Speech or Universal streaming with the current implementation?

[kondricz]

Dear AssemblyAI team,

I have inherited a live transcription piece of code from a previous developer. I have been reading the changelog where Universal-Streaming was introduced.

I am using "assemblyai": "4.9.0" in my frontend code, and I need to know if this one is using Universal Streaming or not?

The official documentation seems to be a code running in backend environment:

https://www.assemblyai.com/docs/speech-to-text/universal-streaming

My questions would be:

• Is my implementation using Universal Streaming or not? (I can't really figure it out from the types or from the documentation)
• Does updating my assemblyai package for 4.13.3 will make it use Universal Streaming? Or would it give an option for that?

Thank You in advance, Just a confused silly dev.

My code:

"use client";

import { useRef, useState, useEffect, useCallback } from "react";
import {
  RecordRTCPromisesHandler as RecordRTC,
  StereoAudioRecorder,
} from "recordrtc";
import { RealtimeTranscriber } from "assemblyai";

const initTranscriber = async () => {
    // Don't initialize if we're already initializing or have an instance
    if (
      isTranscriberConnecting.current ||
      isTranscriberConnected.current ||
      realtimeTranscriber.current
    ) {
      return;
    }

    isTranscriberConnecting.current = true;

    try {
      const token = await getToken();

      realtimeTranscriber.current = new RealtimeTranscriber({
        token,
        sampleRate: 16000,
      });
.... REST OF THE CODE

[Fortelly]

I think the docs could be so much clearer, but according to the guide here you need to connect to an v3 endpoint but the source code for RealtimeTranscriber says:

const defaultRealtimeUrl = "wss://api.assemblyai.com/v2/realtime/ws";

Whereas the source for StreamingTranscriber says:

const defaultStreamingUrl = "wss://streaming.assemblyai.com/v3/ws";

[Shivam-002]

Is there any update on this?