Network tools
Two of these need a server. We say which, and exactly why.
3 tools
In this category
Almost everything on this site runs without touching the network. Two tools in this category cannot, and rather than blur that we have given them their own badge. Your IP address is by definition something only the far end of a connection can tell you, and measuring bandwidth requires something on the other side to send and receive bytes. The user agent parser is the opposite case: it is pure string analysis, so it carries the same local badge as the rest of the site.
What "needs a connection" means here, precisely
Two pages in this category — the IP lookup and the speed test — make requests to endpoints on this same site. Nothing else on the site does, and the third tool here does not either. There is no file input on any of them, so there is no file that could be sent even in principle.
The IP page asks our endpoint one question — what source address did you see this request come from — and displays the answer along with what can be derived from it. The speed test downloads blocks of random bytes to measure download throughput and posts random bytes back to measure upload throughput. Those bytes are generated on the spot and are not derived from anything on your device.
We would rather label this honestly and lose the simplicity of "nothing ever leaves your browser" than stretch that claim until it is no longer checkable. A privacy claim that has quiet exceptions is worse than one with a stated boundary.
Why your IP address is not one number
On a home or office network your device has a private address — usually starting 192.168 or 10 — that is meaningless outside that network. What the rest of the internet sees is your router's public address, assigned by your ISP, and that is what these pages report.
You may well have two public addresses: an IPv4 one, usually shared with other customers through carrier-grade NAT, and an IPv6 one. Which of them a given site sees depends on how your connection and that site are configured, so two IP checkers can legitimately disagree.
Geolocation from an IP address is an estimate derived from registry records, not a measurement. City-level accuracy is often wrong, and mobile connections are frequently placed at the wrong end of the country. Country-level is usually right, which is why streaming and shopping sites use it.
Why a speed test result is a range, not a number
A single measurement tells you the throughput between your device and one server over a few seconds. Change the server, the time of day, the Wi-Fi channel or the distance to your router, and the number changes — none of those changes mean your connection has been altered.
The most common cause of a disappointing result is the local wireless link rather than the internet connection behind it. Testing over Ethernet, or standing next to the router, separates those two possibilities in about a minute.
Latency usually matters more than bandwidth for how fast a connection feels. Video calls, gaming and ordinary web browsing are dominated by round-trip time and by how much that time varies, not by peak megabits. A connection with modest bandwidth and low, stable latency feels faster than the reverse.
What we do not do
We do not keep a record tying your IP address to a visit, we do not build a profile, and we do not embed a third-party measurement network on these pages. Standard server request logs exist at the hosting layer, as they do for every website; that is described on our privacy page.
The endpoints are also explicitly excluded from the offline cache. Caching a personalised response like "your IP address" and later serving it to somebody else would be a privacy incident, so the service worker is written to never store them.
Frequently asked questions
Why do these two tools need a server when nothing else here does?
Because both answers only exist on the network. Your public IP is what a remote server observes, and throughput can only be measured against something that sends and receives data.
Are any of my files sent by these tools?
No. Neither page has a file input. The speed test transfers randomly generated bytes that have nothing to do with anything on your device.
Why does another site show a different IP address for me?
Usually because one of you is reporting IPv6 and the other IPv4, or because a VPN or proxy is applied to some traffic and not other traffic.
Is the location shown for my IP address accurate?
Country level is usually correct. City level is an estimate from registry data and is often wrong, especially on mobile networks.
Why is my measured speed lower than what I pay for?
Advertised figures are maximums under ideal conditions. Wi-Fi, distance from the router, the device itself and normal network contention all reduce what you actually measure.