https://livekit.io logo
Join Slack
Powered by
How to setup egress so I can record user question ...
# ask-ai
w
How to setup egress so I can record user question as separate audio and agent responses as separate audio files?
t
Thanks for reaching out, Chiranjeevi. I'm an AI assistant and I'm here to help. Give me a moment while I research this topic.
If you want to record user and agent audio separately in LiveKit, here are a few ways you can do it: • Dual Channel Audio Mixing: You can use room composite egress with dual channel audio mixing. This records both participants in one file but keeps their audio on separate channels (like agent on the left and user on the right). • Individual Participant Recording: Another option is participant egress, which records each participant’s audio into separate files. This way, their tracks are completely isolated. • Track-Specific Recording: You can also use track composite egress to record specific audio tracks individually—great if you want more precise control over what gets recorded. Just a heads-up: each method needs access to a cloud storage provider (like Amazon S3, Google Cloud Storage, or Azure Blob Storage) to save the recordings. If you’d like, I’m happy to walk you through the steps for any of these or help you get set up! Sources: livekit/livekit_egress.pb.go | Recording Participants | LiveKit Docs | Recording agent sessions | LiveKit Docs
w
Python code for individual participants recording
t
Here’s a handy example of Python code to help you record each participant (both user and agent) separately using LiveKit’s participant egress feature:
Copy code
from livekit import api

async def record_individual_participants():
    # Record the agent
    agent_req = api.ParticipantEgressRequest(
        room_name="your-room-name",  # Swap this out with your actual room name
        identity="agent-identity",  # Use the agent’s identity here
        file_outputs=[api.EncodedFileOutput(
            filepath="agent-recording-{time}",
            s3=api.S3Upload(
                access_key="your-access-key",  # Your S3 access key goes here
                secret="your-secret",         # Your S3 secret key here
                region="your-region",         # And your S3 region
                bucket="your-bucket"          # Don’t forget your S3 bucket name
            )
        )]
    )

    # Record the user
    user_req = api.ParticipantEgressRequest(
        room_name="your-room-name",  # Same here, replace with your room name
        identity="user-identity",   # And the user’s identity
        file_outputs=[api.EncodedFileOutput(
            filepath="user-recording-{time}",
            s3=api.S3Upload(
                access_key="your-access-key",  # Your S3 access key again
                secret="your-secret",         # Your S3 secret key
                region="your-region",         # Your S3 region
                bucket="your-bucket"          # Your S3 bucket name
            )
        )]
    )

    lkapi = api.LiveKitAPI()
    await lkapi.egress.start_participant_egress(agent_req)
    await lkapi.egress.start_participant_egress(user_req)

# Just call this function within your async environment when you’re ready
Just swap out placeholders like 
your-room-name
, 
agent-identity
, 
user-identity
, and your S3 credentials with your real info. If you hit any snags or want me to walk you through it, just give me a shout—I’m here to help!
h
getting this error TwirpError(code=invalid_argument, message=no supported codec is compatible with all outputs, status=400)
t
It looks like the error might be related to a codec compatibility issue. To help you better, could you please share a bit more about your setup? For example: • Which audio codec you’re using (like Opus, AAC, or others) • The output format and how you’ve set up storage (such as file type or S3 settings) • Any special configurations or customizations you’ve made in your egress request With these details, I’ll be able to understand the issue more clearly and guide you toward the best solution. Thanks for sharing!
h
audio codec is ogg storage is s3 not custom configurations
t
I don't have the answer you're looking for. You could also try asking your question: • in one of the other Slack channels or • to https://deepwiki.com/livekit/livekit_composite which is trained on all LiveKit source code If you find the answer, please post it here to help others!
Open in Slack
PreviousNext
Powered by