How much bandwidth is your connection actually delivering?
Measures your download and upload bandwidth via our endpoint. Run it first without a VPN, then with, to quantify the overhead.
Three steps, no install.
Latency measurement
4 HEAD requests to our endpoint, median of the 2 central values. Gives the real round-trip latency.
Download test
Downloads a random 10 MB blob. Throughput is calculated on total reception time.
Upload test
POSTs a 5 MB payload. Throughput is calculated on full transmission time.
Methodology
We download an incompressible binary payload from our Railway server, then we POST an equivalent payload back. Latency is measured via a HEAD request first. Accuracy: ±10% on residential connections, to be compared with an official Speedtest.
A well-configured VPN (WireGuard / NordLynx protocol) typically incurs 5-15% bandwidth loss and +10 to +40 ms of latency depending on server distance. Beyond that, the server or protocol is poorly chosen.
Everything you need to know.
What loss is normal with a VPN?
Between 5 and 15% on a well-configured VPN using WireGuard or NordLynx, on a geographically close server. Beyond 30%, the server is overloaded or the protocol unsuitable.
Why does my measurement vary so much between two tests?
Network variance is strong (peak hours, Wi-Fi, ISP congestion). For a reliable measurement, run 3 tests on the same server and keep the median.
Does the test consume my mobile data?
Yes, it downloads ~10 MB and sends ~5 MB. On a limited 4G plan, use sparingly. The result remains valid on Wi-Fi (latency and throughput are identical).
Why does my official Speedtest show a different rate?
The tools don't use the same servers or measurement durations. Our test measures to our Railway server, not the nearest Speedtest server. Compare with the same tool over time, not between tools.
You've seen the gap. Close it.
30-day money-back guarantee. No card required for tools.