Internet speed test
Measure your connection's download and upload throughput and latency against our server. Runs in your browser: no plugins, nothing to install.
Download
-
Mbps
Upload
-
Mbps
Latency
-
ms
How to use the internet speed test
- Open this page: the test starts automatically.
- Wait for the download and upload phases to finish.
- Read your download, upload, and latency; run it a few times for a reliable figure.
What the numbers mean
Download and upload are throughput in megabits per second (Mbps: divide by 8 for megabytes); latency is the round-trip delay in milliseconds. The test runs against vpn.golf's server in New Jersey, so it approximates your raw line speed before any VPN: it isn't measuring through an exit node yet. Results vary with Wi-Fi vs wired, distance, ISP plan, and congestion, so run it a few times. A VPN typically costs a little speed in exchange for privacy. To check responsiveness to a specific host, use the ping test.
Code & API examples
Use this from the command line or your code.
Download throughput (curl)
curl -o /dev/null -w '%{speed_download} bytes/s\n' "https://vpn.golf/api/tools/speedtest/download?bytes=10000000"See all endpoints at /api/tools/.
Frequently asked questions
Download is how fast data comes to you, upload is how fast it goes out, and latency (ping) is the round-trip delay before data starts moving. Download and upload are throughput; latency is responsiveness, which matters most for calls and gaming.
Mbps is megabits per second, the standard way ISPs and this test report speed. MBps is megabytes per second, what you see when a file downloads. Divide Mbps by 8 to get MBps, so 100 Mbps is about 12.5 MBps.
This test measures the path to vpn.golf's single server in New Jersey, while other tools pick a nearby server to flatter the number. Distance, your route to us, and current load all shift the result, so treat it as one honest data point.
Not yet. It measures your raw connection to our control-plane server, which does not route VPN traffic. It approximates the speed your line can deliver before any tunnel. Exit-node speed testing comes later as the hole fleet rolls out.
Usually a little. Encryption and the detour through an exit node add some overhead and distance, so you trade a slice of speed for privacy. On a well-provisioned, nearby exit the cost is small and rarely noticeable for browsing or streaming.
Most home and mobile plans are asymmetric by design, giving you far more download than upload because people consume more than they send. A result where upload is a fraction of download is normal, not a fault.
It depends on use. Roughly 25 Mbps handles HD streaming, 100 Mbps comfortably covers a busy household, and gigabit is plenty for almost anything. For calls and gaming, low and stable latency matters more than raw throughput.
Wi-Fi versus a wired connection, distance and interference, how many devices share the line, your ISP plan, congestion at peak hours, and the route to our server. Wired and off-peak runs give the most consistent numbers.
A single run can be skewed by a momentary spike, a background download, or transient congestion. Running it a few times and looking at the range gives a truer picture than one number, which may be a fluke high or low.
Wi-Fi loses speed to distance, walls, interference, and older hardware, so the wireless result is often well below what the line delivers. To test your actual plan, connect by Ethernet and run it again.
Yes. Many networks slow during evening peak hours when neighbors are all online and shared infrastructure gets congested. If your speeds dip predictably at night, congestion is the likely cause rather than your own equipment.
It can hint at it. If general speeds look fine here but specific services like streaming or torrents are far slower, your ISP may be throttling those by type. Comparing this raw speed against in-app speeds is the simplest tell.
Want to hide your IP for real?
vpn.golf is a no-logs WireGuard VPN. Pick a hole, take the shot.
Step up to the tee for free