LiveKit + PeerMetrics

LiveKit monitoring for WebRTC apps

Monitor LiveKit call quality, packet loss, jitter, and SFU behavior across every session.

Add your LiveKit Room to PeerMetrics to capture what users experienced, label SFU nodes, and debug call issues faster without changing your normal room flow.

PeerMetrics SDK

Quick setup

  • Initialize PeerMetrics for the user and conference.
  • Register the LiveKit SDK integration.
  • Keep your normal call setup flow.

How it works

  1. Initialize PeerMetrics before or alongside your room setup.
  2. Create or reuse your LiveKit Room instance.
  3. Call addSdkIntegration() with the Room object.
  4. Optionally provide serverId and serverName to identify SFU nodes in dashboards.
Example
import { Room } from 'livekit-client'

const room = new Room(roomOptions)

const peerMetrics = new PeerMetrics({
  apiKey: 'YOUR_API_KEY',
  userId: currentUser.id,
  conferenceId: roomName
})

peerMetrics.addSdkIntegration({
  livekit: {
    room,
    serverId: 'livekit-sfo-1',
    serverName: 'LiveKit SFU - SFO'
  }
})

Based on the PeerMetrics SDK integration docs.

Voice AI on LiveKit

Monitor voice AI sessions with the same LiveKit evidence

If you are building realtime voice AI on LiveKit, PeerMetrics helps you inspect the media-quality side of each agent session, including packet loss, jitter, reconnections, and the SFU path behind a bad interaction.

That makes it easier to separate model or turn-taking issues from actual network, browser, device, or transport problems when users say the assistant sounded delayed, robotic, or dropped out.

Use the same session evidence across engineering, support, and operations to debug production voice experiences faster.

Built for production teams

See what happened in every LiveKit room

Track room quality, label SFU nodes, and give support and engineering the same session evidence when users report call issues.

Start with the open-source SDK or get help rolling it out faster if you want dashboards, ingestion, and instrumentation in place sooner.

See the LiveKit demo Talk to us about LiveKit monitoring View LiveKit integration docs

Monitor call quality with session-level evidence

Give engineering, support, and operations the facts they need to understand what happened in every LiveKit session.
  • Session evidence for LiveKit products
    Replace guesswork with session evidence

    Building a LiveKit product is already difficult. WebRTC adds network, device, and browser behavior that is hard to reason about from user reports alone.

    Give your team the call-level evidence they need to improve quality, isolate regressions, and resolve incidents faster.

  • Developer relaxing
    Spot quality patterns before they become recurring support issues

    Network conditions, devices, and client behavior change constantly. Monitor the patterns that degrade LiveKit calls before they become recurring support problems or engineering fire drills.

  • Happy support team
    Help support and engineering work from the same facts

    Your support team should not have to guess what happened. Share the same call diagnostics across support, engineering, and operations so handoffs stay fast and credible.