We're building a voice authentication system where new users call our VAPI assistant, provide their details, and record their voice for authentication using Resemblyzer. The issue we're facing is that we cannot access real-time audio streaming through WebSocket during the call - we only receive the complete audio after the call ends. We need live voice data streaming to capture and store it in S3 immediately for authentication purposes. Has anyone successfully implemented real-time audio capture with VAPI, or are there specific WebSocket events/configurations we should be using? i Searched lot but not get the best Answer. https://cdn.discordapp.com/attachments/1430149244874002453/1430149245226320045/vapi.png?ex=68f8b9c6&is=68f76846&hm=65d5ce546912cad5150f35a6efbff4e872f165a645883db47c33ec8b3c9cec29&

To capture real-time audio during a Vapi call, you must use the WebSocket transport. After initiating a call with the WebSocket transport, connect to the provided

websocketCallUrl

and stream audio as it arrives. Here’s a Node.js example:

js
const WebSocket = require('ws');
let pcmBuffer = Buffer.alloc(0);
const ws = new WebSocket('wss://api.vapi.ai/YOUR_CALL_ID/transport');
ws.on('message', (data, isBinary) => {
  if (isBinary) {
    pcmBuffer = Buffer.concat([pcmBuffer, data]);
    // Upload pcmBuffer to S3 here or in chunks
  }
});

This lets you process and store audio in real time—no need to wait for the call to end. Be sure to initiate the call with the WebSocket transport to get the correct URL and streaming behavior. For more details, see the [Vapi WebSocket transport documentation](https://docs.vapi.ai/calls/websocket-transport). Source: - [WebSocket Transport](https://docs.vapi.ai/calls/websocket-transport)

Hi Huzaifa Tahir, If you have questions about implementing WebSocket transport with Vapi for streaming audio or any other use case, here are a few guidelines based on your needs: 1. **WebSocket Audio Streaming**: - Vapi supports real-time, bidirectional audio communication via WebSockets. For configuration, you can set up calls in either PCM or Mu-Law formats. Check [here](https://docs.vapi.ai/calls/websocket-transport) for detailed instructions. 2. **Integration Steps**: - Initiate a WebSocket call by making an API request with your assistant ID and desired audio format. - Utilize the provided WebSocket URL to establish a connection and handle real-time audio data. You can send and receive both binary audio and JSON control messages. 3. **Error Handling**: - Implement error handling within your WebSocket connections to manage events such as connection errors or closed connections. For more info consult [this guide](https://docs.vapi.ai/calls/websocket-transport). If you need further clarity on configuration or specific examples, feel free to ask!