NETWORK GUIDES
How to test Wi-Fi and mesh network performance
Compare wired and wireless performance, room-by-room coverage and mesh paths without confusing signal bars with measured backhaul quality.
Start with a repeatable baseline
Use the same device and browser for each location. Record whether the device is wired or wireless, the room and approximate distance from the access point. Keep the test profile, duration and data allowance the same. Repeat a wired baseline where available so the comparison has a reference.
Signal bars do not describe the whole path
A device can have a strong connection to a nearby mesh satellite while that satellite has a weak or congested backhaul to the main router. Conversely, weak client signal can reduce performance even when the backhaul is healthy. End-to-end browser results alone cannot separate these cases.
Compare locations without changing several things at once
Test near the main router, near a satellite and in the problem room. Repeat each location under similar household activity. Note if the device roams between access points. Use your own labels: this website does not discover which satellite or radio a browser is associated with.
What this browser can actually see
The tester measures application transfers and browser-observed timing to its server. The optional local browser check observes main-thread responsiveness and coarse browser-provided hardware hints. It does not read radio retries, channel utilisation, NIC negotiation, satellite parent links or actual OS CPU percentages.
A comparison record you can reproduce
Record three runs at each position: wired baseline where available, near the main access point, near a satellite and in the problem room. For each run keep the date and time, your own location label, connection type, browser/device, duration, data allowance, download and upload Mbps, idle and directional loaded RTT, sample validity and household activity. Export each result before changing one factor. Compare the middle result and spread, not only the fastest run. Leave unavailable hardware fields blank; a browser comparison does not discover the mesh topology.
What requires an integration
Direct router and mesh diagnosis needs authorised controller or companion telemetry. Those integrations are not connected in this public beta. The interface marks router, mesh and ISP infrastructure observations as unavailable rather than fabricating a topology. A list of hypothetical devices is not a network scan.
Make a useful before-and-after record
Change one factor at a time, such as device location or using a cable instead of Wi-Fi. Repeat the same measurement and compare results, errors and sample validity. Keep the original exports and note the circumstances. Several repeatable observations are more useful than one unusually fast or slow run.