I want to record both the room egress and just the sound tracks. both recording should be saved

Thanks for reaching out, Anne. I'm an AI assistant and I'm here to help. Give me a moment while I research this topic.

yes please how do I run both? Currently I create a room like so room_request = livekit_api.CreateRoomRequest( name=room_name, empty_timeout=600, # 10 minutes max_participants=10, metadata=metadata_json, egress=livekit_api.RoomEgress( room=livekit_api.RoomCompositeEgressRequest( file_outputs=[ livekit_api.EncodedFileOutput( filepath=f"{room_name}.mp4", s3=livekit_api.S3Upload( access_key=config.s3_upload_access_key, secret=config.s3_upload_secret, bucket=metadata.interview_recording_bucket, region="us-east-1", ), ) ] ) ), )

Hey there! If you want to run both room composite egress and track egress at the same time in LiveKit, you’ll just need to make separate API calls for each one. Here’s a simple way to do it using Python:

import asyncio

# Start room composite egress
room_egress_task = egress_service.start_room_composite_egress(
    RoomCompositeEgressRequest(
        room_name="my-room",
        file_outputs=[EncodedFileOutput(filepath="room-{time}.mp4")]
    )
)

# Start track egress
track_egress_task = egress_service.start_track_egress(
    TrackEgressRequest(
        room_name="my-room",
        track_id="track-id", 
        file=DirectFileOutput(filepath="track-{time}.mp4")
    )
)

# Wait for both to start
room_info, track_info = await asyncio.gather(
    room_egress_task, 
    track_egress_task
)

Just a heads-up: each egress will have its own unique ID, and since they run independently, make sure your server can handle both at once. If you hit any bumps or want to chat more about this, I’m here to help! Sources: pkg/service/egress.go | packages/livekit-server-sdk/src/EgressClient.ts | src/test/kotlin/io/livekit/server/EgressServiceClientTest.kt | livekit-api/livekit/api/egress_service.py | protobufs/livekit_egress.proto

how do I know the "track-id"

Hey there! If you’re looking to find the "track-id" for a specific audio or video track in LiveKit, here are a few easy ways to do it: • From TrackPublication: If you have the track publication handy, just grab the

sid

field—that’s the unique ID for the track. For example:

trackId = trackPublication.sid

. • Server API: You can also use the LiveKit server API to list all participants and their tracks. This will give you the track IDs for every track in the room. • By Name or Source: If you know the track’s name or source, you can find the track publication and then get its

sid

. If you want, I’m happy to walk you through any of these methods step-by-step—just let me know! Sources: Universal Egress | libwebrtc/src/media_stream.rs | findMicroTrackId | LiveKit Agents | lib/src/types/track_identifier.dart | Core API primitives | LiveKit Docs