How it works
Exchanges keep their servers in specific data centers. The usual path from your network to them often takes a long way through other networks' nodes. PingNitro sends your terminal's traffic over a short dedicated route straight to the exchange servers.
Exchanges only
Only addresses from a closed list of exchanges go through the route. The rest of your internet goes direct.
We can't see your data
Your terminal's connection to the exchange is encrypted (TLS). We can't access your API keys or orders.
A backup entry
The app picks the best entry and switches to the backup one in 10 seconds if something fails (in our tests).
Measurements by exchange
BTC trade-feed latency, median, from our own measurements.
| Exchange | Direct | Via PingNitro | Difference |
|---|---|---|---|
| Binance Futures | 144 ms | 78 ms | −66 |
| Binance Spot | 165 ms | 72 ms | −93 |
| OKX | 147 ms | 79 ms | −68 |
| Bitget | 153 ms | 91 ms | −62 |
| Gate | 204 ms | 75 ms | −129 |
| Bybit, BingX | 106 ms | — | direct |
Bybit and BingX run their servers in Singapore — the direct path to them is shorter, so the app sends them directly, outside the route.
Gain by city
The gain depends on where you connect from. Below is our estimate by city; for your own connection, run the latency test.
Estimate for other cities
| Gain of more than 100 ms | Saint Petersburg, Minsk, Almaty, Yerevan |
| Gain of 60–90 ms | Moscow, Yekaterinburg, Istanbul, Tbilisi, Baku |
| Almost no gain | Novosibirsk and cities further east |
This is an estimate, not a feed measurement: TCP connection time to a Binance server directly vs. time to our entry point plus the link to the exit (the Globalping service); median of 1–3 probes per city, a single observation. For your own connection, the latency test is more accurate.