https://livekit.io logo
Join Slack
Powered by
# github
  • m
  • m
    new commit pushed to 
    <https://github.com/livekit/livekit/tree/master|master>
    by boks1971 
    <https://github.com/livekit/livekit/commit/2f002388b642c656dc82f1f095d0c8d44e81424c|2f002388>
    - Use participant close reason in remove. (#3639) livekit/livekit
  • m
  • m
    new commit pushed to 
    <https://github.com/livekit/livekit/tree/master|master>
    by boks1971 
    <https://github.com/livekit/livekit/commit/8d3902af3647e9985064bfffb8354bb73af537f7|8d3902af>
    - Protocol to pick up cgroups v2 memory path fix (#3640) livekit/livekit
  • m
  • m
    new commit pushed to 
    <https://github.com/livekit/livekit/tree/master|master>
    by boks1971 
    <https://github.com/livekit/livekit/commit/a1f4e88edbd46bc3c12fef3fc0ab56fac2d3328e|a1f4e88e>
    - Update protocol to latest, got bit by tag (#3641) livekit/livekit
  • m
  • m
    new commit pushed to 
    <https://github.com/livekit/livekit/tree/master|master>
    by boks1971 
    <https://github.com/livekit/livekit/commit/b83190a3790bc5e46b6d1a6d6ce499af7b1e952d|b83190a3>
    - protocol update to fix memory stats path (#3642) livekit/livekit
  • m
    #690 Ensure audio frame when publishing Pull request opened by hiroshihorie Ensure at least one audio frame was generated before returning publish(). livekit/client-sdk-swift
  • m
    #569 Make some audio config extensions public Pull request opened by bcherry This makes it easier to implement a custom configure function by copying the default one, which relies on these non-public extensions to function. Now you can drop in the default and the only thing to change is the log lines. livekit/client-sdk-swift
    • 1
    • 1
  • m
    • 1
    • 2
  • m
    new commit pushed to 
    <https://github.com/livekit/livekit/tree/master|master>
    by boks1971 
    <https://github.com/livekit/livekit/commit/d4ab1142cba9cc5f61d96954c8c60ef60373521f|d4ab1142>
    - Redact address (#3643) livekit/livekit
  • m
    #3644 Where I can pull request to livekit docs? Issue created by gusye1234 Hi there. Livekit is really great! I'm building a open-source long-term memory project called Memobase, which recently add example of how to use livekit and Memobase to build Voice Agent that can remember. The example is using the latest version of livekit's Agent SDK (can be founded here), and I noticed that the mem0(another long-term memory project) example is outdated in your official docs. I want to add this example in livekit's docs for users to have a start point to build personal voice agent, but the livekit-docs repo is archived. I'm wondering where I can pull request to livekit docs? livekit/livekit
  • m
    #3645 Distributed LiveKit: Agent worker can only serve the node it registered on Issue created by umarniz ## Summary In a 3-node LiveKit cluster, a single LiveKit Agent worker registers successfully on Node C. When a new room request lands on Node B, the server cannot hand that job to the worker on Node C. Instead it 1. Tries (and fails) to assign the job on Node B. 2. Falls back to Node A, which creates the room without any Agent worker attached. Result: the room starts, but the Agent never connects and hence there is only silence. --- ## Topology
    Copy code
    Node A ── LiveKit server
Node B ── LiveKit server              (receives the room request)
Node C ── LiveKit server              (worker registers here)
    --- ## Logs ## Node A
    Copy code
    2025-05-06T05:42:55.411Z  INFO  livekit  service/roomallocator.go:164  selected node for room  {"room":"<room_name>","selectedNodeID":"<livekit_node_id>"}
2025-05-06T05:42:55.423Z  INFO  livekit.api  service/twirp.go:124  API RoomService.CreateRoom  {"service":"RoomService","method":"CreateRoom","room":"<room_name>","request":{"name":"<room_name>","metadata":"<metadata>"},"duration":"16.212838ms","status":"200"}
2025-05-06T05:42:55.596Z  INFO  livekit.webhook  webhook/url_notifier.go:124  sent webhook  {"event":"room_started","id":"<event_id>","webhookTime":1746510175,"room":"<room_name>","roomID":"<room_id>","url":"<webhook_url>","queueDuration":"42.041µs","sendDuration":"173.904564ms"}
2025-05-06T05:42:55.646Z  INFO  livekit  agent/client.go:161  failed to send job request  {"error":"no workers with sufficient capacity","namespace":"","jobType":"JT_ROOM","agentName":""}
2025-05-06T05:42:56.493Z  INFO  livekit  service/roommanager.go:405  starting RTC session  {"room":"<room_name>","roomID":"<room_id>","participant":"<participant_id>","pID":"<process_id>","remote":false,"nodeID":"<livekit_node_id>","numParticipants":0,"participantInit":{"Identity":"<participant_id>","Client":{"sdk":"<sdk>","version":"<sdk_version>","os":"<os>","osVersion":"<os_version>","deviceModel":"<device_model>","browser":"<browser>","browserVersion":"<browser_version>"}}}
2025-05-06T05:42:56.496Z  INFO  livekit  service/roommanager.go:938  created TURN password  {"username":"<turn_username>","password":"<turn_pass>"}
2025-05-06T05:42:56.808Z  INFO  livekit.transport  rtc/transport.go:546  ice reconnected or switched pair  {"room":"<room_name>","roomID":"<room_id>","participant":"<participant_id>","pID":"<process_id>","remote":false,"transport":"PUBLISHER","existingPair":{"localProtocol":"udp","localCandidateType":"host","localAddress":"<local_ip>","localPort":"<local_port>","remoteProtocol":"udp","remoteCandidateType":"prflx","remoteAddress":"<remote_ip>","remotePort":"<remote_port>"},"newPair":{"localProtocol":"udp","localCandidateType":"host","localAddress":"<local_ipv6>","localPort":"<local_ipv6_port>","remoteProtocol":"udp","remoteCandidateType":"host","remoteAddress":"<remote_ipv6>","remotePort":"<remote_ipv6_port>"}}}
2025-05-06T05:42:56.922Z  INFO  livekit.pub  rtc/participant.go:1826  mediaTrack published  {"room":"<room_name>","roomID":"<room_id>","participant":"<participant_id>","pID":"<process_id>","remote":false,"kind":"audio","trackID":"<track_id>","webrtcTrackID":"<webrtc_track_id>","rid":"","SSRC":"<ssrc>","mime":"audio/red","trackInfo":{"sid":"<track_id>","type":"AUDIO","source":"MICROPHONE","mimeType":"audio/red","mid":"1","stream":"camera"},"fromSdp":true}
2025-05-06T05:42:56.957Z  INFO  livekit.webhook  webhook/url_notifier.go:124  sent webhook  {"event":"track_published","id":"<event_id>","webhookTime":1346510176,"room":"<room_name>","roomID":"<room_id>","participant":"<participant_id>","pID":"<process_id>","url":"<webhook_url>","queueDuration":"14.513µs","sendDuration":"32.059281ms"}
2025-05-06T05:42:57.704Z  INFO  livekit  rtc/room.go:469  participant active  {"room":"<room_name>","roomID":"<room_id>","participant":"<participant_id>","pID":"<process_id>","remote":false,"publisherCandidates":"<redacted>","subscriberCandidates":"<redacted>","connectionType":"udp"}
2025-05-06T05:42:57.744Z  INFO  livekit.webhook  webhook/url_notifier.go:124  sent webhook  {"event":"participant_joined","id":"<event_id>","webhookTime":1346510177,"room":"<room_name>","roomID":"<room_id>","participant":"<participant_id>","pID":"<process_id>","url":"<webhook_url>","queueDuration":"34.951µs","sendDuration":"39.824442ms"}
    ### Node B
    Copy code
    2025-05-06T05:42:55.632Z	WARN	livekit.agents	service/agentservice.go:401	failed to assign job to worker	{"jobID": "<job_id>", "namespace": "", "agentName": "", "jobType": "JT_ROOM", "room": "<room_name>", "roomID": "<room_id>", "workerID": "<OLD_WORKER_ID_1>", "retry": true, "error": "worker not available"}
2025-05-06T05:42:55.639Z	WARN	livekit.agents	service/agentservice.go:401	failed to assign job to worker	{"jobID": "<job_id>", "namespace": "", "agentName": "", "jobType": "JT_ROOM", "room": "<room_name>", "roomID": "<room_id>", "workerID": "<OLD_WORKER_ID_2>", "retry": true, "error": "worker not available"}
2025-05-06T05:42:55.644Z	WARN	livekit.agents	service/agentservice.go:401	failed to assign job to worker	{"jobID": "<job_id>", "namespace": "", "agentName": "", "jobType": "JT_ROOM", "room": "<room_name>", "roomID": "<room_id>", "workerID": "<OLD_WORKER_ID_3>", "retry": true, "error": "worker not available"}
2025-05-06T05:42:55.644Z	WARN	livekit.agents	service/agentservice.go:391	no worker available to handle job	{"jobID": "<job_id>", "namespace": "", "agentName": "", "jobType": "JT_ROOM", "room": "<room_name>", "roomID": "<room_id>", "error": "no workers with sufficient capacity"}
    ### Node C
    Copy code
    2025-05-06T05:42:44.567Z	INFO	livekit.agents	service/agentservice.go:292	worker registered	{"namespace": "", "jobType": "JT_ROOM", "agentName": "", "workerID": "<CORRECT_WORKED_ID>"}
    --- ## Expected behavior • Node B should be able to dispatch the 
    JT_ROOM
    job to the worker already registered on Node C, or the cluster should transparently proxy the request. ## Actual behavior • Node B fails to find any available worker and eventually the allocator selects Node A, which creates the room without an Agent worker. --- ## Reproduction steps 1. Run three LiveKit nodes (A, B, C) backed by the same Redis/state store (running in sentinel with one master node). 2. Start one Agent worker only on Node C (job type 
    JT_ROOM
    ). 3. Send a 
    CreateRoom
    request that the allocator routes to Node B. 4. Observe worker-not-available warnings on Node B and room creation on Node A. --- ## Environment | Component | Version | | -------------- | --------------------------------------- | | LiveKit server | v2.8.0 | | LiveKit Agent | v0.x (worker) | | Redis | 7.2 | | Deployment | Docker Compose (one container per node) | | OS | Ubuntu 22.04 | --- ## Question Why can’t Node B delegate the job to the registered worker on Node C? Is worker discovery limited to the local node by design, or is there an extra configuration flag required to make workers globally visible across the cluster? Any insight or pointers would be greatly appreciated! livekit/livekit
  • m
    #691 Fix external webcam visibility on macOS Pull request opened by pblazej Handles 
    .unspecified
    position livekit/client-sdk-swift
    • 1
    • 1
  • m
    #685 Pre-connect final API Pull request opened by pblazej on <!date^1745926330^{date_short}|2025-04-29T11:32:10Z> • Implements agent changes: 
    Track
    vs 
    Room
    attributes • Keeps 
    LocalTrack
    identity while recording/publishing (required by the above ⬆️) • Adds misc checks for robustness (e.g. don't send to non-agent participants) • Adds new scoped API for easier cleanup on errors livekit/client-sdk-swift
    • 1
    • 1
  • m
    #693 Extract agent components Pull request opened by pblazej Moves/tweaks a few scattered pieces of agent logic from the voice assistant example livekit/client-sdk-swift
    • 1
    • 1
  • m
    #2703 Occasional bug, after livekit has been running for a week, room creation error: could not find any available nodes Issue created by keaipeng123 on <!date^1714963212^{date_short}|2024-05-06T02:40:12Z> [image](https://private-user-images.githubusercontent.com/90554992/328068654-7153c207-320b-43b4-b815-b33d3a4e06d4.png?jwt=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.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.Xr7OBMDRmkrUsr1W5vn0Hz0OwROCPbNbVOsYevXvnr8) livekit version:v1.5.3 Program code if _, err = core.RoomClient.CreateRoom(context.Background(), &livekit.CreateRoomRequest{ Name: rn, EmptyTimeout: 10 * 60, // 10min MaxParticipants: 10, Metadata: owner + "-" + skill, }) livekit/livekit
    • 1
    • 1
  • m
    #692 Echo agent tests Pull request opened by hiroshihorie on <!date^1746537747^{date_short}|2025-05-06T13:22:27Z> WIP, needs echo agent to be running. livekit/client-sdk-swift
    • 1
    • 1
  • m
  • m
    1 new commit pushed to 
    <https://github.com/livekit/livekit/tree/master|master>
    by boks1971 
    <https://github.com/livekit/livekit/commit/2fff36cb35364e6cccea3df2e46d902cb9944e82|2fff36cb>
    - Stub MoveParticipant so that cloud can include the latest protocol. (#3646) livekit/livekit
  • m
    #3647 Process Terminated with SIGILL (Exit Code -4) When User Starts Speaking Issue created by Divyansh-adda247 # Bug Report: Process Terminated with SIGILL (Exit Code -4) When User Starts Speaking ## Issue Summary When a user begins speaking into the microphone, a supervised process is terminated with exit code -4 (SIGILL - Illegal Instruction), causing audio processing to fail. This crash occurs before any UserStateChangeEvent is triggered, suggesting the failure happens early in the audio processing pipeline. ## Environment • Operating System: Ubuntu 24.04 • Python Version: 3.12.3 • LiveKit Package Version: 1.0.2 • Hardware: x86_64 ## Steps to Reproduce 1. Initialize the LiveKit application 2. Connect to a voice/audio session 3. Start speaking into the microphone 4. Observe the process crash with exit code -4 ## Error Details
    Copy code
    process exited with non-zero exit code -4
    This error is triggered when the supervised process terminates with SIGILL (signal 4), which indicates an illegal instruction was executed. The error happens in 
    supervised_proc.py
    around line 280-283: if self._exitcode != 0 and not self._kill_sent: logger.error( f"process exited with non-zero exit code {self.exitcode}", extra=self.logging_extra(), ) ## Technical Analysis The error appears to be related to the Foreign Function Interface (FFI) used for audio processing. When audio input is detected from the microphone: 1. Audio data is captured and passed to the audio processing modules 2. This data is processed through 
    AudioProcessingModule
    in 
    apm.py
    3. The module calls native code via FFI in 
    _ffi_client.py
    4. When attempting to execute CPU instructions in the native code, a SIGILL is triggered 5. The supervised process is terminated before any user state change events can be emitted The most likely cause is a CPU architecture mismatch between the compiled native FFI library (`liblivekit_ffi.so`/`.dylib`/`.dll`) and the current hardware, or the use of CPU instructions not supported by the current processor. ## Impact • Audio processing fails completely when users attempt to speak • Speech-to-text functionality is unavailable • Real-time voice communication is broken ## Additional Information The error only occurs when audio input is first processed, suggesting it's related to the initialization of audio processing components or the first execution of specialized CPU instructions for audio processing. ## Possible Solutions to Investigate 1. Provide compatible builds of the LiveKit FFI libraries for different CPU architectures 2. Add runtime detection of CPU capabilities and use fallback implementations when specific instructions aren't available 3. Set the 
    LIVEKIT_LIB_PATH
    environment variable to point to a compatible library version 4. Implement more robust error handling in the audio processing pipeline to prevent crashes ## Logs and Diagnostics Agent state changed: from thinking to speaking Metrics collected type='metrics_collected' metrics=TTSMetrics(type='tts_metrics', label='livekit.plugins.google.tts.TTS', request_id='06a4e98e246e', timestamp=1746639797.2195292, ttfb=0.22073470099712722, duration=0.2774173969955882, audio_duration=2.112833333333335, cancelled=False, characters_count=37, streamed=False, speech_id='speech_ccc7eb5cfab0') 2025-05-07 174317,474 - httpcore.http11 - DEBUG - receive_response_body.complete 2025-05-07 174317,474 - httpcore.http11 - DEBUG - response_closed.started 2025-05-07 174317,474 - httpcore.http11 - DEBUG - response_closed.complete Metrics collected type='metrics_collected' metrics=LLMMetrics(type='llm_metrics', label='livekit.plugins.openai.llm.LLM', request_id='chatcmpl-BUd5YRsHyNdWIy7pl3OJbeMX7uc7V', timestamp=1746639797.4746995, duration=1.7644396450050408, ttft=1.0211968020012137, cancelled=False, completion_tokens=62, prompt_tokens=2076, prompt_cached_tokens=0, total_tokens=2138, tokens_per_second=35.13863462290481, speech_id='speech_ccc7eb5cfab0') Metrics collected type='metrics_collected' metrics=TTSMetrics(type='tts_metrics', label='livekit.plugins.google.tts.TTS', request_id='ec7f88af0a11', timestamp=1746639797.9792652, ttfb=0.40468507799960207, duration=0.7448762679996435, audio_duration=7.512416666666596, cancelled=False, characters_count=156, streamed=False, speech_id='speech_ccc7eb5cfab0') Metrics collected type='metrics_collected' metrics=TTSMetrics(type='tts_metrics', label='livekit.plugins.google.tts.TTS', request_id='66994be7d3e4', timestamp=1746639798.5305398, ttfb=0.24503060399729293, duration=0.5509810129951802, audio_duration=5.972833333333295, cancelled=False, characters_count=116, streamed=False, speech_id='speech_ccc7eb5cfab0') Agent state changed: from speaking to listening Chat History updated ["assistant: Hello Himesh, it's great to meet you! Before we dive into the technical questions, could you please tell me a bit about what motivated you to pursue a degree in Computer Science and Engineering? \n\nAdditionally, what have been some of your most memorable experiences or achievements during your time at university?"] Metrics collected type='metrics_collected' metrics=STTMetrics(type='stt_metrics', label='livekit.plugins.deepgram.stt.STT', request_id='', timestamp=1746639812.839311, duration=0.0, audio_duration=0.05, streamed=True) Metrics collected type='metrics_collected' metrics=VADMetrics(type='vad_metrics', label='livekit.plugins.silero.vad.VAD', timestamp=1746639813.7896044, idle_time=18.08289429100114, inference_duration_total=0.01760541099793045, inference_count=32) Metrics collected type='metrics_collected' metrics=VADMetrics(type='vad_metrics', label='livekit.plugins.silero.vad.VAD', timestamp=1746639814.8101447, idle_time=19.103434784003184, inference_duration_total=0.01616592599748401, inference_count=32) 2025-05-07 174336,281 - livekit.agents - ERROR - process exited with non-zero exit code -4 2025-05-07 174336,281 - ERROR livekit.agents - process exited with non-zero exit code -4 {"pid": 37275, "job_id": "AJ_XTg747wkhjik"} livekit/livekit
  • m
    #3648 Add Moving participant to another room Pull request opened by cnderrauber it is implemented in cloud only since the destination room can exist in different node with the source room livekit/livekit
    • 1
    • 6
  • m
    1 new commit pushed to 
    <https://github.com/livekit/livekit/tree/master|master>
    by cnderrauber 
    <https://github.com/livekit/livekit/commit/793b383a5292b23df0a1e79fcd989b0af404b8fa|793b383a>
    - Add Moving participant to another room (#3648) livekit/livekit
  • m
    #3556 feat: add opentelemetry instrumentation for tracing Issue created by marcklingen on <!date^1743010199^{date_short}|2025-03-26T17:29:59Z> ## Context I'm one of the maintainers of github.com/langfuse/langfuse. livekit is growing in popularity and I've gotten multiple requests by community members who'd like to get traces of their livekit applications in order to evaluate them. ## Suggestion I'd suggest to follow the example of the AI SDK (docs) by adding native Opentelemetry-based instrumentation to livekit. Thereby, the implementation is vendor-agnostic and teams can use an LLM-specific backend like Langfuse (tracing, evals, prompt management) or their traditional observability platforms (Grafana, Datadog, ...). Potential addition next to traces: Exporting metrics (livekit docs) via OpenTelemetry would allow teams to easily track them within their observability stack. ## Resources • https://opentelemetry.io/blog/2025/ai-agent-observability/ • https://opentelemetry.io/docs/specs/semconv/gen-ai/ • https://langfuse.com/docs/opentelemetry/get-started • https://github.com/orgs/langfuse/discussions/5235 livekit/livekit
    • 1
    • 1
  • m
    #3649 Event conversation_item_added only adds Agent chat items when using Gemini Realtime API Issue created by alinalihassan Describe the bug Following the guide in the docs (https://docs.livekit.io/agents/build/events/) for events while building an AI Agent, I wanted to log the interaction between the user and the agent. Using 
    @session.on("conversation_item_added")
    , I've noticed that the user messages are not caught by the event. Server • Version: 1.0.6 • Environment: local • any other information about your deployment setup Client • SDK: Agents Playground & Local console • Version: [0.x.x] To Reproduce Steps to reproduce the behavior: 1. Set up Gemini Realtime model (google.beta.realtime.RealtimeModel) 2. Add an event handler for session for "conversation_item_added" 3. Check logs and see only agent events, no user events Expected behavior I expected the user messages to be invoked as well, as it's mentioned in the documentation by this: "A ConversationItemAddedEvent is emitted when a item is committed to the chat history. This event is emitted for both user and agent items.". Additional context Code:
    Copy code
    import logging
import os

from dotenv import load_dotenv
from livekit import agents
from livekit.agents import (
    Agent,
    AgentSession,
    AutoSubscribe,
    ConversationItemAddedEvent,
    JobContext,
    RoomInputOptions,
)
from livekit.agents.llm import AudioContent, ImageContent
from livekit.plugins import google, noise_cancellation
from mem0 import AsyncMemoryClient

# Setup logging
logger = logging.getLogger("agent")

load_dotenv()

# Initialize Mem0 client
memory_client = AsyncMemoryClient(api_key=os.getenv("MEM0_API_KEY"))

# Voice configurations
INTRO_VOICE = "Leda"  # Friendly, welcoming voice
TUTOR_VOICE = "Puck"  # More authoritative, teaching voice


class TestAssistant(Agent):
    def __init__(self) -> None:
        super().__init__(instructions="You are a helpful voice AI assistant.")


async def entrypoint(ctx: JobContext):
    <http://logger.info|logger.info>("Starting entrypoint function")

    try:
        await ctx.connect(auto_subscribe=AutoSubscribe.AUDIO_ONLY)
        <http://logger.info|logger.info>("Connected to room successfully")
        await ctx.wait_for_participant()

        session = AgentSession(
            llm=google.beta.realtime.RealtimeModel(
                model="gemini-2.0-flash-exp",
                voice=INTRO_VOICE,
                temperature=1.0,
            ),
        )

        @session.on("conversation_item_added")
        def on_conversation_item_added(event: ConversationItemAddedEvent):
            print(
                f"Conversation item added from {event.item.role}: {event.item.text_content}. interrupted: {event.item.interrupted}"
            )

        # Start with the IntroAgent
        await session.start(
            room=ctx.room,
            agent=TestAssistant(),
            room_input_options=RoomInputOptions(
                noise_cancellation=noise_cancellation.BVC()
            ),
        )

    except Exception as e:
        logger.error(f"Error in agent setup or execution: {e}", exc_info=True)


if __name__ == "__main__":
    agents.cli.run_app(agents.WorkerOptions(entrypoint_fnc=entrypoint))
    livekit/livekit
  • m
  • m
    1 new commit pushed to 
    <https://github.com/livekit/livekit/tree/master|master>
    by paulwe 
    <https://github.com/livekit/livekit/commit/aee34ffefad89f135378bf11b8461e866d83e01b|aee34ffe>
    - log request for agent dispatch api (#3650) livekit/livekit
  • m
    #3651 Prometheus Metrics for auth Issue created by wrenix counter metric for passing / failing auth request. i like to know if a high request of invalid token happen without parsing / grep for
    Copy code
    2025-05-08T15:45:29.935Z    WARN    livekit    service/auth.go:95    error handling request    {"status": 401, "method": "GET", "path": "/rt
c/validate", "error": "invalid token: eyJhbGciOiJIUzI1<redact-shortted>MF00w, error: go-jose/go-jose: err
or in cryptographic primitive"}
    livekit/livekit
  • m
    #3652 Prometheus Metrics for http Issue created by wrenix It would be nice to see the default metrics of http request like: https://github.com/prometheus/client_golang/blob/v1.22.0/examples/middleware/httpmiddleware/httpmiddleware.go Not only the metrics on the PrometheusRequest:
    Copy code
    # HELP promhttp_metric_handler_requests_in_flight Current number of scrapes being served.
# TYPE promhttp_metric_handler_requests_in_flight gauge
promhttp_metric_handler_requests_in_flight 1
# HELP promhttp_metric_handler_requests_total Total number of scrapes by HTTP status code.
# TYPE promhttp_metric_handler_requests_total counter
promhttp_metric_handler_requests_total{code="200"} 206
promhttp_metric_handler_requests_total{code="500"} 0
promhttp_metric_handler_requests_total{code="503"} 0
    livekit/livekit