The quickest way to test internet speed is in the browser: open the 404 VPN speed test, press "Measure speed" and wait about ten seconds. You get three numbers: download speed in megabits per second, ping and jitter in milliseconds. Nothing to install, no account, and it works on a phone as well as a laptop.
A single run, though, is not enough to draw conclusions. This guide explains what each number means, how to measure so the result is fair, why two services show different figures for the same line, and what to do when the speed is well below what your plan promises.
What a speed test measures
Download speed. How many megabits per second reach you from the server. It is the number that matters for video, web pages, updates and file downloads. The speed test downloads test data for a few seconds and divides the amount received by the time it took.
Ping. The time a short request needs to reach the server and come back, in milliseconds. Ping is not about downloading fast but about responding fast: it decides how snappy calls, games and page loads with dozens of small requests feel.
Jitter. How much the ping varies between consecutive requests. If one reply takes 20 ms and the next 80 ms, the connection is unsteady. Downloads do not care, but voice and video do: you hear stutter and see the picture fall apart. Microsoft's network requirements for Teams treat jitter up to 30 ms as acceptable.
No upload test. Ours is deliberately light and has no server-side component. If you need to check upload speed, for example before streaming, use a service with a full upload test.
How to test internet speed properly
- Stop everything that downloads. Updates, torrents, cloud sync and a video in another tab all share the line with the test.
- Use a cable if you can. Wi-Fi through two walls easily halves the speed. No cable? Move next to the router.
- Run it two or three times. One run can land on a burst of traffic; look at the typical result, not the best one.
- Compare morning and evening. In the evening your provider's link is shared by everyone in the building, and a dip at that time is normal rather than a fault.
- Check phone and laptop separately. An old phone or Wi-Fi adapter can be the bottleneck on its own.
Why different tests disagree
Speedtest, Fast.com and our tool rarely agree to the megabit, and that is not a bug. Four things differ.
- Where the server is. Our server is in Amsterdam; the farther you are, the higher the ping. Services with a server network pick the nearest one, sometimes in your own city.
- How many streams. Big tests download over several parallel connections and squeeze the line to its limit. Ours uses a single stream, like an ordinary file download in a browser.
- How long. In the first seconds a connection is still ramping up (that is how TCP works), so a short test reports less than a long one.
- What is reported. Some show a peak, some an average. Our test shows the average over the whole run and the current speed while it runs.
For the question "can my line handle HD video and calls" any test will do. To compare two connections with each other, use the same test for both.
Megabits and megabytes
Providers sell speed in megabits per second (Mbps); browsers and messengers show downloads in megabytes per second (MB/s). A byte has eight bits, so a "100 Mbps" plan means roughly 12.5 MB/s. If a 1 GB file takes about a minute and a half, you are getting an honest hundred megabits.
Speed with and without a VPN
A VPN adds one more hop on the path, so a small drop in speed and a slightly higher ping are normal. To see how much you actually lose, run the test twice in a row: without the VPN and with it, on the same device and the same network. If the difference is several times over, the cause is usually the server or the protocol rather than the VPN itself; the breakdown and a fix-it checklist are in does a VPN slow down your internet. On networks that throttle UDP, switching to VLESS over TCP helps; you can check your network with the UDP block test.
When the speed is below your plan
- Restart the router. Obvious, but it fixes the problem more often than anything else: it clears accumulated errors and moves to a less crowded Wi-Fi channel.
- Check the cable and the port. An old patch cord or a 100 Mbps port on the router will not pass a gigabit plan.
- Change the Wi-Fi band. 5 GHz is faster but weaker through walls; 2.4 GHz reaches farther but is crowded with neighbours.
- Compare over cable. If the speed matches the plan over cable but not over Wi-Fi, the problem is the router or interference, not the provider.
- Collect results and contact the provider. Three or four cabled measurements at different times of day are a solid argument for support.
In short
- Three numbers: download, ping, jitter. Download is for video and files; ping and jitter are for calls and games.
- Measure over cable, with no background downloads, two or three times, morning and evening.
- Tests disagree because of server location, streams and duration; compare connections with one and the same test.
- With a VPN, measure twice: without it and with it.
404 VPN runs on VLESS and WireGuard, picks the fastest server automatically, and shows latency and load for every server before you connect. How the connection works is on the how it works page; you can get started on the home page.