Skip to content

Cookies: only the ones needed.

This site remembers your language, your theme and whatever you decide here. The rest — the visitor counter and the YouTube videos — only loads if you say so. Read the cookie policy.

Antonio CintoraSISTEMAS
TOOL

Speed test

Measure what your connection actually delivers right now: download, upload, ping, and — the part almost nobody shows — how much the line chokes when it is busy.

The test takes about 35 seconds and moves real data: it measures for 15 seconds in each direction, which on fast fibre can reach 2 GB across download and upload. On mobile data, that comes out of your allowance. The bytes are served by Cloudflare's network, so Cloudflare sees your IP address, exactly as it would if you opened any site of theirs.

0
Mbps
 
ms
PING
ms
JITTER
Mbps
DOWNLOAD
Mbps
UPLOAD
THE TEST
YOUR IP
MEASURED AGAINST
PROTOCOL
DATA USED

You will not see your city or your ISP here, and that is not an oversight: those values do travel in the response, but the browser will not let a site on another domain read them. Finding them out would mean handing your IP to a third party, which is exactly what this tool does not do. Your IP is shown to you, on your screen, and never leaves your browser.

TO MAKE SENSE OF THIS

Why is it lower than what I pay for?

Almost always Wi-Fi. A 1000 Mbps line behind a 2.4 GHz radio realistically delivers 60 to 100 Mbps, and that is the last few metres failing, not the line: try a cable before you call anyone. Then there is protocol overhead. TCP, IP and TLS eat 2 to 3 per cent, so the most an application can ever see on a 1000 Mbps line is around 940. And finally the computer itself: at those speeds the browser can be the bottleneck.

Why do you not measure packet loss?

Because a browser cannot, and we would rather not show a made-up number. Measuring real loss means sending individual UDP packets and counting the ones that never arrive, and a browser has no access to that. What other tools label packet loss is usually something else wearing the name.

What is jitter and when does it matter?

It is how much the ping varies from one measurement to the next. For downloading it is irrelevant, because throughput is what counts. For video calls and gaming it matters more than raw speed, because high jitter is what breaks up the audio no matter how much bandwidth you have.

What is the chart for?

To show what the average hides. A line that holds flat at 600 and one that starts at 900 and collapses to 300 after ten seconds produce the same average, and they are nothing like the same connection. If the curve drops off a step partway through, that is usually your ISP throttling you after an initial burst, and it is exactly the kind of thing worth having captured before you ring customer support.

Why does the result change between runs?

Because it measures the network as it is at that moment, and the network changes. Time of day, whoever else is using the line, Wi-Fi conditions and load on the far end all count. If you want to compare, run it a few times at different moments and look at the set rather than one figure.

Why don't you show my ISP, like other tools do?

Because knowing it would mean asking someone, and asking would mean handing over your IP address. Cloudflare does send that value in the response, but browser security rules stop a site on another domain from reading it, so technically it never reaches this page. I could run a server in the middle to resolve it; I don't, because then your IP would pass through my hands, and I would rather it didn't.

Do you store my results?

No. There is no database, no result link and no record of who measured what. Your IP is shown to you, on your screen, and never reaches a server of ours. If you want to keep a measurement, the copy button gives you the text to paste wherever you like.

None of this is stored. The measurement happens between your browser and Cloudflare's network, which serves the bytes; neither the figure nor your address stays here.