Frequently asked questions
Straight answers about the test, the numbers, privacy and what a browser can and cannot measure.
Using the speed test
Does the speed test start automatically?
Yes, once the page has painted and the browser is idle. It skips the automatic start in Data Saver mode, in a background tab, or if you already ran a test in this browser tab — then Test again appears instead.
What is Single vs Multi connections?
Single measures with one TCP stream. Multi uses four parallel streams so a fast line can be filled more fully. Neither reports Wi-Fi versus Ethernet — browsers usually do not expose that — so the control focuses on something that actually changes the measurement.
How do I switch between Kbps, Mbps and Gbps?
Use Speed unit under the dial. The gauge scale, centre readout and download/upload cards update together. Your preference is saved on this device.
How much data does a test use?
It varies with your speed, so treat any figure as an estimate. The engine caps a full run at about 100 MB in total — 60 MB download and 40 MB upload — and a slower line transfers less because the stage ends on its time budget first. Latency-only tools use a few hundred kilobytes; the bufferbloat test uses the most because it saturates the line twice.
Why does my phone show a different result from my laptop?
Different Wi-Fi radios, processors and power settings. The slower device often caps the measurement. Compare like with like, or test both over Ethernet.
Why is the result lower than my plan?
Wi-Fi, device limits, VPNs, background traffic and peak-hour congestion all reduce the measured figure. Test over Ethernet, try Multi connections, then contact your provider if the wired result is still far below the plan.
Can I stop a test?
Yes. Every test has a Stop control; the run is cancelled and no result is reported for the cancelled direction.
Accuracy
How accurate is the measurement?
Within roughly 5–10% on a stable connection with a capable device. That is the practical limit of browser-based testing, not a limitation specific to this site.
Why do different speed tests disagree?
Different endpoints, different ramp-up handling, single versus multiple streams and different test durations. This site lets you choose Single or Multi so you can compare both styles. See the methodology page for the rest of our method.
Is the ping the same as a game ping?
No. It is the round-trip time to the nearest measurement edge, including browser and HTTP overhead. In-game ping adds the route to the game server.
Do you show results when the test fails?
No. Failed measurements show an explanation and a retry option. Placeholder or invented numbers are never displayed.
Privacy and data
Do you store my results?
No. Results are calculated in your browser and, if you share one, encoded in the link itself.
Do you see my IP address?
The measurement endpoint sees it, as any web request does. This site does not log it or associate it with results.
Are cookies used?
No cookies. Theme, speed-unit and connection-mode preferences use local storage on your device only.
Tools and limitations
Can a browser measure packet loss?
Not the UDP packet loss a game or call experiences. Our packet loss page measures failed HTTP requests and labels it as request loss.
Why can I not choose a test server?
Requests are routed to the nearest edge automatically, which keeps the path representative. A manual picker would let you measure a route your traffic never takes.
Does the site work offline or without JavaScript?
Pages render without JavaScript, but measurement requires it. Without JavaScript the tools explain that they need it rather than showing empty results.