I am working on a workflow to transcribe live audio from a Twilio call using Deepgram via n8n. What I’ve done so far: 1.Set up Twilio to stream call audio using <Stream> with a WebSocket proxy. 2. My proxy forwards media.payload as base64 audio to my n8n webhook. 3. Using a Function node in n8n to decode base64 and send to Deepgram for transcription. The problem: My Decode Caller Audio node in n8n does not output any data, and Deepgram receives...
Originally asked by Maddy Khan on Stack Exchange (stackoverflow) (CC BY-SA 4.0).