I am trying to build realtime voice translation react-native application using mediasoup. My doubt is how I can pass the audio stream of webrtc in mediasoup server to audio translation pipeline made using services like openai whisper and Google cloud services.