Our idea — not observed on the original site. This lab compares browser signals with a selected network profile. The actual local connection has no public-IP intelligence. Modeled inputs and recorded proxy measurements are labeled.
| Signal | Result | Reason |
|---|
Ready. External probes are off until you select the public check.
IP addresses do not have a clock. We compare device UTC with server UTC, then compare the browser's UTC offset with the IP region's expected offset at the same instant. Daylight saving and equivalent timezone names are handled by offsets.
The default WebRTC probe gathers host ICE candidates only, without camera or microphone access. Browsers often hide addresses behind mDNS. The optional public probe adds a STUN server and an HTTPS egress lookup. Neither mode can prove the absence of leaks. An unavailable result remains unknown. An IPv6 candidate is not compared against an IPv4 baseline, and a private LAN address alone is not treated as a proxy leak.
Datacenter, VPN, proxy and Tor labels come from the selected dataset or recorded provider estimate. They are not inferred from the browser. Selecting a recorded proxy does not route this browser through it.