GEARCHECK LABObserve how downloaded bytes arrive over time from the GearCheck Lab server, with a maximum 32 MiB test payload.
READY — choose an action below
No measurement yet.
The graph groups bytes delivered to JavaScript into half-second intervals. Interval Mbps is received payload bits divided by the actual elapsed interval, not the maximum speed of your internet package. Browser buffering can make bytes arrive in bursts. The run ends after 15 seconds or the 32 MiB cap, whichever comes first. A very fast connection may reach the cap before many intervals are collected; that is shown as an insufficient time span rather than a fabricated stability grade. Partial results remain visible after Stop.
Server load, HTTP connection reuse, Wi-Fi, VPNs and browser buffering influence the result. It is not raw packet-loss measurement, a guarantee of sustained line speed or a nearest-server test. Test payload and protocol overhead consume data; actual network usage can be slightly higher than payload. Stop and hiding the page abort pending requests. Only generated server test bytes are downloaded; no local files are uploaded.
Device previews, recordings, selected files and report notes stay in your browser. Downloaded files remain on your device. Network tests send generated probes to GearCheck Lab. DNS lookup sends your requested public domain name and type to our server and the Cloudflare public resolver only after you start a query. Local recordings, drawings, files, device names and battery logs are not uploaded. Ordinary page requests and usage measurement are described in our privacy policy.