Skip to main content
All Tools

WebRTC & P2P Speed Test

Measure upload, download and latency between two devices over a WebRTC peer-to-peer (P2P) connection. Open the test in both browsers and share the joining code, with nothing to install. Use it before sending large files or to compare devices and networks. The local test runs both peers on one device and measures browser overhead only, not your network or internet speed. Upload and download run one after the other, with the same reliable, unordered bulk channel and flow control used for file transfers.

Two-Device Test

Measure real P2P speed between two devices. One starts the test, the other joins with a code.

Local Test

Quick single-device test. Measures DataChannel overhead, not real network speed.

How to run a P2P speed test

  1. On the first device, choose Two-Device Test. You'll get a short code, a link and a QR code.
  2. On the second device, scan the QR code or open the link. It can be a phone, a laptop on another network, or a friend's computer on the other side of the world.
  3. Keep both tabs open while the test connects, measures latency, then sends 50 MiB in each direction, one direction at a time. When it's done, share the results link or compare against a TURN relay using the same byte count.

If the two devices won't connect at all, the step-by-step guide to testing WebRTC connections walks through finding out why.

Two-device test vs local test

The two-device test is the real one. Data crosses the actual network between the two devices over a WebRTC DataChannel, using the same 14 KiB chunks, reliable unordered delivery and buffer flow control SpectrShare uses for file transfers. A separate ordered channel carries test commands. Speeds are measured on the receiving side.

The local test runs both peers in one tab and pushes 200 MiB in each direction through a loopback connection. No data leaves your machine, so it tells you how fast your browser and CPU can move DataChannel traffic, not how fast your network is. Both peers compete for the same tab's processing time. A two-device test distributes that work across devices and may be faster or slower, so the local result isn't a ceiling for network transfers.

What the results mean

Upload is Device A to Device B and download is B to A, so a big gap between them usually points at a difference in the two paths or devices. These are sequential measurements, not simultaneous bidirectional throughput. The average speed divides all delivered bits by elapsed transfer time; min, max and stability describe short sampling intervals. Latency is the round-trip time for a small message, with min, average, max and jitter. The 1 GB estimate uses the average of the two directions and excludes file reads, hashing and disk writes. A local estimate doesn't predict a transfer to another device.

Connection type matters most. Direct P2P means the devices talk to each other and you'll avoid a relay hop. Via TURN Relay means a server is forwarding every byte, so the result reflects that server's bandwidth as much as yours. If you keep getting relayed, check your NAT type with the WebRTC connection test.

Don't expect this number to match your ISP speed test. That tool measures a short hop to a server built to be fast. A WebRTC speed test measures a path between two ordinary devices, limited by whichever side is slower.

P2P speed test FAQ

Does this measure upload and download at the same time?

No. Upload and download run sequentially, with a separate result for each direction. The local test runs both peers in one browser tab, so CPU load and browser scheduling affect its speed. It measures browser DataChannel throughput and does not set a ceiling for transfers between two devices.

Is this P2P speed test different from a normal internet speed test?

Yes. A normal speed test measures your connection to a nearby server picked for speed. This P2P speed test measures the connection between two specific devices, which depends on both internet connections, both NATs and firewalls, and the route between them. It's the number that matters when you send a file directly to someone.

Do both devices need to install anything?

No. Both sides run the test in a browser tab. One device starts the test and gets a code and QR code; the other opens the link or enters the code to join.

Why is my peer-to-peer speed lower than my ISP speed?

A P2P transfer can only go as fast as the slower side allows, and usually that limit is upload speed. If one device has 100 Mbps and the other has 10 Mbps upload, expect roughly 10 Mbps. Extra network hops, Wi-Fi and a TURN relay all take more off the top.

What does "Via TURN Relay" mean in my results?

It means the devices couldn't connect directly, usually because of a strict NAT or corporate firewall, so a relay server forwarded the data. Relay speed is limited by the server's bandwidth and adds latency. About 85% of WebRTC connections manage a direct path.

How can I get faster P2P transfers?

Put both devices on the same network if you can, since LAN transfers skip the internet entirely and can reach hundreds of Mbps. Use a wired connection on at least one side, and pause streaming or large downloads while you test.

Related tools: TURN server tester, WebRTC connection test, SpectrShare connectivity check