Research Question: do two Chromium-based browsers on the same machine and the same network produce identical WebRTC results? We ran NetRiskScan's own WebRTC Leak Test in both, back to back, and they didn't quite match.
| Test date | August 13, 2026 |
| Environment | Windows 11, single machine, single network connection, run sequentially seconds apart |
| Browsers | Google Chrome 151.0.7922.109 · Microsoft Edge 151.0.4129.78 |
| Method | Automated (Playwright), headless — see Limitations |
| Tool | NetRiskScan WebRTC Leak Test, live production site |
The results, side by side
| Chrome 151 | Edge 151 | |
|---|---|---|
| ICE candidate collection | Succeeded | Partially succeeded |
| STUN servers succeeded | 3 / 3 | 2 / 3 |
| Total candidates gathered | 11 | 10 |
| Public candidates | 5 | 4 |
| NetRiskScan's own STUN server | Success (191.222.208.203, 8839 ms) | No public address returned |
| Cloudflare STUN server | Success (289 ms) | Success (248 ms) |
| Google STUN server | Success (287 ms) | Success (248 ms) |
| Leak verdict | No obvious leak | No obvious leak |
In both browsers, every server-reflexive candidate that did come back matched the same real HTTP exit address (191.222.208.203) — so neither run found an actual leak. The difference is reliability, not privacy: our own STUN server answered Chrome in under 9 seconds and returned nothing usable to Edge in the same test run, while the two third-party STUN servers (Cloudflare, Google) answered both browsers quickly and consistently.
Why this is worth noting
Chrome and Edge share the same underlying Chromium engine and, in this test, an identical major version number (151) — a reasonable person could assume identical networking behavior follows from that. This result shows that assumption doesn't fully hold: browser-specific differences (timing, connection handling, or Edge's additional enterprise-oriented network stack — see Microsoft Edge and WebRTC Privacy) can produce a real, measurable difference in which STUN servers successfully respond, even with the rendering engine held constant.
Limitations
This is a single test run on a single machine and network, using browser automation (Playwright) in headless mode rather than a manual interactive session — headless execution can behave differently from a normal user session in ways specific to networking timeouts, and this result should be read as "what this automated run observed," not a general claim about Chrome vs. Edge WebRTC reliability. We do not have Firefox installed on this machine, so it's excluded from this round rather than estimated. A single STUN server timing out once is also not necessarily reproducible — network conditions vary run to run, and we did not repeat the test multiple times to establish a failure rate.
Conclusion
Run the WebRTC Leak Test in whichever browser you actually use day to day — this result is a concrete demonstration of why "I tested it in Chrome" doesn't necessarily generalize to every browser on the same machine, even closely related ones.