Free WebRTC tool

Connection quality analyzer

Run a live WebRTC connection and watch bitrate, round trip time, jitter, packet loss and MOS score graph in real time. Add a TURN server to measure your real network path.

// Media source
Synthetic (no permission) Real camera and mic
// Path mode
Local loopback Force TURN relay (real network)

Synthetic mode generates a moving test pattern and a tone, so no camera or microphone permission is needed.

Estimated MOS
--
Measuring...
Collecting the first few seconds of statistics.
local loopback
Round trip
--ms
Jitter
--ms
Packet loss
--%
Bitrate
--kbps
Resolution
--
Codec
--
Bitrate (kbps)--
Round trip time (ms)--
Jitter (ms)--
Packet loss (%)--
// Sender and receiver
local (sending)
remote (received)

What this measures, and what it does not

This analyzer builds a real WebRTC peer connection between two RTCPeerConnection objects in your browser, sends live media across it, and polls getStats once per second to graph the results. It is the same data your own WebRTC application would see, without writing any code.

Local loopback mode

In loopback mode both peers live in this tab, so media never leaves your machine. Round trip time will be near zero and packet loss will usually be zero. That is expected and it is not a bug. Loopback is useful for checking codec negotiation, encoder performance, resolution and frame rate adaptation, and how your device copes under CPU load. It cannot tell you anything about your network.

TURN relay mode

To measure your real network you need media to actually leave your machine. Relay mode does this by forcing iceTransportPolicy to relay against a TURN server you supply. Every packet then travels from your browser out to the TURN server and back, so the round trip time, jitter and loss you see are genuine measurements of the path between you and that server. This is the honest way to test real network quality from a browser, and it doubles as a load test for your own TURN deployment.

How the MOS score works

MOS, or Mean Opinion Score, rates call quality from 1 (unusable) to 5 (perfect). This tool estimates it using a simplified ITU E-model that combines round trip time, jitter and packet loss into an R factor, then converts that to a MOS value. Above 4.0 is good, 3.6 to 4.0 is acceptable, below 3.6 users start complaining, and below 3.0 a call is difficult to hold. It is an estimate, not a substitute for real listening tests, but it tracks real quality well.

This analyzer is part of Sazrika Foundation. Also try the ICE Candidate Inspector to check your NAT and TURN reachability, the SDP Analyzer to decode offers and answers, and the Media Device Tester to check your camera and microphone.