Ping Tool
Measure latency, jitter and packet loss to any server - from our network, not yours
100% free. This tool needs our server to do the work, so what you submit is sent there, processed and then discarded - it is never stored.
15 of 15 ping tests left today.
About this tool
Measure how quickly a server answers, and how consistently. This ping test runs from our network rather than your browser, so it tells you about the host rather than about your own broadband. It reports minimum, average and maximum round-trip time, jitter (how much the latency varies between probes) and packet loss. ICMP echo is used wherever the environment allows raw sockets; when it does not, latency is measured by timing a TCP handshake instead, and the result tells you which method produced the numbers rather than presenting one as the other.
Diagnose a slow or unreachable website
Work outwards: is the host answering, is the page slow, and is the certificate still valid.
- 1Ping the hostLatency, jitter and packet loss.You are here
- 2Run a speed testReal browser-rendered performance audit.Open
- 3Check the certificateAn expired one blocks every visitor.Open
- 4Inspect DNSConfirm the records point where you expect.Open
How to use this tool
- Enter a hostType a hostname or IP address. A full URL is fine too - the host part is used.
- Choose how many probesBetween 1 and 10. More probes give a better read on jitter and intermittent loss.
- Read the resultsAverage latency tells you how far away the server is; jitter tells you how stable the path is; packet loss above zero points at a real problem.
Key features
- Runs server-side, so a slow result means the host is slow rather than your own connection.
- Jitter is reported, not just average latency - that is what makes calls and games feel choppy.
- Falls back to a TCP handshake when ICMP is blocked, and says so, instead of silently reporting nothing.
- A per-probe breakdown shows whether a bad average came from one outlier or from consistent slowness.
- Chooseable TCP port, so you can probe the service that matters rather than just the host.
- Free, no signup and no software to install.
Common uses
- Checking whether a website is reachable from outside your own network.
- Comparing latency to two hosting providers before choosing one.
- Diagnosing a laggy game or video call by looking at jitter rather than raw speed.
- Confirming that a firewall change did or did not open a port.
- Spotting intermittent packet loss that a single test would miss.
- Sanity-checking a monitoring alert before escalating it.
Tips for better results
- Run more probes when you suspect intermittent loss - a single probe cannot show it.
- High jitter with low average latency usually points at congestion, not distance.
- Choose the port your service actually uses so the TCP fallback tests something meaningful.
- Latency to a CDN-backed site reflects the nearest edge, not the origin server.
- A refused connection still proves the host is up and answering, unlike a timeout.
- Compare against a known-good host to tell a network problem from a host problem.
Mistakes to avoid
- Reading this as a test of your own internet speed - it measures our network to the host.
- Treating no reply as proof a site is down; plenty of servers drop ICMP deliberately.
- Comparing an ICMP number against a TCP number as if they were the same measurement.
- Judging stability from one probe rather than watching jitter across several.
- Pinging a hostname behind a CDN and concluding something about the origin server.
Frequently asked questions
No. The probes run from our server, so the result describes the path from our network to the host. That is what you want when checking whether a site is slow for everyone or only for you - run it alongside a local test to tell the two apart.
ICMP echo needs raw socket permission, which most container environments do not grant, and many hosts drop ICMP anyway. When that happens we time a TCP handshake to the chosen port instead. It is a real latency measurement, just typically a few milliseconds higher than an ICMP reply.
Jitter is the average variation between consecutive replies. A connection with 40 ms latency and 2 ms jitter feels smooth; one with 40 ms latency and 30 ms jitter feels choppy on calls and in games, even though the average looks identical.
Some probes got no reply within the timeout. Occasional loss on a busy path can be normal, but consistent loss points at congestion, a failing link or a firewall dropping traffic. Run more probes to see whether it is steady or intermittent.
No. Private, loopback and link-local addresses are refused, because a public tool that probes internal networks is a port scanner. Use a local ping command for addresses inside your own network.