client.swift

socket.on("server-text-to-speech") {data, ack in
    print("server-text-to-speech")

    // The data structure for this message type is here:
    // https://github.com/descriptinc/descript-web-v2/blob/staging/pkg-js/brain-buddy-protocol/src/ServerSentEvents.ts#L27-L48
    if let firstItem = data.first as? [String: Any],
        guard let progress = firstItem["progress"] as? String else {
            return
        }
         
        if progress == "first" {
          // First packet of text to speech
        }
         
        // All messages except ones with progress=='final' have an audioChunk associated
        if progress != "final" {
            let bytes = firstItem["audioChunk"] as? Data { // Assuming 'audioChunk' is of type Data
            self.aacBuffer.append(bytes);
        }
        
        if progress == 'final' {
          // No more text to speech audio is coming.
        }       
    } else {
        print("Key not found in dictionary, or data format not as expected.")
    }
}