Enter a host (and optional port) to measure the TCP round-trip time from our network — a quick reachability and latency check. Defaults to port 443.
How to use the ping / latency test (tcp)
Enter a hostname (and a port, or leave it at 443).
Press Test.
Read the min, average, and max round-trip time across the samples.
Reading the results
This opens several TCP connections to the host and times how long each takes to establish — a round-trip latency measurement that works even where ICMP "ping" is blocked, and that confirms the actual service port is reachable. The numbers are measured from vpn.golf's network to the host, not from your own device, so they're best for comparing how reachable a server is rather than measuring your personal connection. Under 30 ms is excellent, 30–100 ms is fine for most uses, and a wide gap between min and max means high jitter. To check whether a specific port is open at all, use the port checker.
Code & API examples
Use this from the command line or your code. The API is free, GET-only, and returns JSON.
Latency is the time a signal takes to travel to a host and back, measured in milliseconds. This test opens TCP connections to a host and reports the round-trip time, so lower numbers mean a snappier connection.
Classic ping uses ICMP echo packets. This tool instead times how long a TCP connection takes to establish, which works where ICMP is blocked and reflects whether the actual service port is reachable.
ICMP requires raw socket access that browsers don't grant for security reasons. So this test runs server-side from vpn.golf and uses TCP connect timing, which is the closest reliable measure available over the web.
From vpn.golf's network to the target host, not from your own device. It tells you how reachable and fast a server is from our datacenter — useful for comparing hosts, but not a substitute for testing from your machine.
Under 30 ms feels instant, 30-100 ms is fine for browsing and most gaming, 100-250 ms is noticeable, and above 300 ms feels sluggish. Distance to the host sets the floor — you can't beat the speed of light.
Physical distance, the number of network hops, congestion, the host's own load, and the connection type (fiber is lower than satellite). A nearby, lightly loaded server over wired internet gives the lowest times.
Jitter is the variation between successive round-trip times. The gap between this test's min and max hints at it. High jitter disrupts real-time uses like video calls and gaming even when average latency looks fine.
The host didn't complete a TCP connection in time — it may be down, the port may be closed or firewalled, or the name may not resolve. Try a port you know is open, like 443, to confirm reachability.
It defaults to 443 because almost every web host accepts connections there. Pick the port of the service you actually care about — 80, 22, or a game server's port — to measure latency to that specific service.
A single measurement can be skewed by a momentary hiccup or a cold connection. Repeating it and reporting min/avg/max reveals the typical latency and how stable the path is, which one sample can't show.
Routing through a VPN adds a hop, so latency usually rises a little, especially to a distant exit. The hit is small to a nearby hole and is often worth it; sometimes a VPN even lowers latency by avoiding ISP congestion or throttling.
Test the host from a different network to isolate whether the problem is your connection or the host, switch from Wi-Fi to wired, close bandwidth-heavy apps, and pick a geographically closer server when you have the choice.