Home/Speed test

Throughput

Downloads an actual asset and reports what arrived. Not a synthetic random-bytes endpoint: the number worth having is the one you get from the same file a player requests, through the same cache, under the same compression rules.

Pulls /media/sustained/sustained-1080p.mp4, roughly 190 MB. The result updates while the transfer is running and is final when it stops.

Reading the number

A single measurement from one browser tab is a lower bound, not a link capacity. TCP spends the first moments ramping up, so a short transfer reports less than the connection can sustain — which is exactly why this test uses the five-minute asset rather than one of the sixty-second clips.

Each run appends a cache-busting query parameter. Without it the second run measures your browser's disk cache and reports a number that has nothing to do with the network.

From the command line

curl reports the same thing without a browser in the way:

curl -o /dev/null -w 'avg %{speed_download} B/s in %{time_total}s\n' \
  https://vid.thisaintu.sbs/media/sustained/sustained-1080p.mp4

For parallel behaviour, which is closer to how a player with several connections behaves:

seq 4 | xargs -P4 -I{} curl -s -o /dev/null \
  -w '%{speed_download}\n' \
  https://vid.thisaintu.sbs/media/sustained/sustained-1080p.mp4?p={}

Please keep automated loops reasonable. There is no rate limit and no key, and that stays true only while nobody treats it as free bandwidth for its own sake.