how do i receive bytes of output audio from my TTS...
# ask-aic
crooked-tiger-31253
06/27/2025, 11:48 PMhow do i receive bytes of output audio from my TTS in livekit’s agent session?
t
tall-airline-16756
06/27/2025, 11:48 PMThanks for reaching out, Faraz. I'm an AI assistant and I'm here to help. Give me a moment while I research this topic.
tall-airline-16756
06/27/2025, 11:48 PMTo receive bytes of output audio from your TTS in a LiveKit agent session, the process typically involves the following steps:
1. The TTS service makes an API request to a provider (e.g., Neuphonic, Hume, Groq, Inworld, or Resemble).
2. The response contains audio data, often base64 encoded.
3. The audio data is decoded into bytes when needed.
4. The bytes are pushed to an
output_emitterpush()output_emitter.flush() Copy code
audio_bytes = base64.b64decode(parsed_message["data"]["audio"])
output_emitter.push(audio_bytes) Copy code
async for data, _ in resp.content.iter_chunks():
output_emitter.push(data)AudioEmitterc
crooked-tiger-31253
06/28/2025, 12:04 AMhow does this work for the elevenlabs plugin? what about the cartesia plugin?
t
tall-airline-16756
06/28/2025, 12:05 AMHey there! Let me break down how the ElevenLabs and Cartesia plugins handle audio output in a super friendly way. 😊 For the ElevenLabs plugin, here's how the magic happens:
1. The plugin connects to the ElevenLabs API and receives audio data in base64 encoded format. Think of it like a digital translator for audio!
2. The audio data gets decoded into bytes using this neat little code: `if data.get("audio"):
b64data = base64.b64decode(data["audio"])
output_emitter.push(b64data)`It's basically turning the encoded message into something our system can understand.
3. The decoded bytes are then sent to an
output_emitter