Benchmarks
Five runs, timed on the clock
Every number on this page came out of an actual compression through the live page on 30 October 2026. Nothing here is projected or averaged from someone else's hardware.
Method
The page was opened in Chromium on Windows 11, a file was handed to the file input, the target size was set, and the clock ran from the click on the compress button until a result existed. One run per file, not a best-of-five. Timing therefore includes reading the file, decoding, encoding and muxing — everything a visitor experiences.
The machine is a small cloud instance with 2 vCPU. That matters: encoding speed scales with cores and with the hardware encoder the browser picks, so a modern desktop will be faster than this table and an older laptop slower.
The runs
| Input | In | Target | Out | Time | Path |
|---|---|---|---|---|---|
| MP4, 1920×1080, 30 fps, 30.44 s | 20.7 MB | 10 MB | 9.81 MB | 40.1 s | Re-encoded to 720p |
| MP4, 1920×1080, 30 fps, 30.44 s | 26.2 MB | 25 MB | 24.4 MB | 31.5 s | Re-encoded, resolution kept |
| MP4, 1280×720, 30 fps, 6.0 s, no audio | 3.9 MB | 2 MB | 1.96 MB | 2.3 s | Re-encoded |
| WEBM, VP9, 854×480, 30 fps, 8.0 s | 873 KB | 1 MB | 906 KB | 7.8 s | VP9 cannot be copied → re-encoded |
| MKV, H.264, 854×480, 30 fps, 8.0 s | 7.96 MB | 10 MB | 7.96 MB | 0.5 s | Already fits → repackaged, streams copied |
What the table shows
- The fast path is about 80× faster. The MKV that already fit was repackaged in 0.5 s; the comparable re-encode took tens of seconds. If your file already fits the cap, you are not paying for an encode at all.
- Time tracks pixels and duration, not file size. The 30-second 1080p clip and the 6-second 720p clip differ by 20× in wall clock because one has roughly 12× the frames and 2.25× the pixels per frame.
- Overshoot is not zero. A 1 MB target produced 906 KB — 9% under, which is the budget working as designed. A 10 MB target produced 9.81 MB. Files land a few percent under the cap rather than exactly on it.
- Repackaging can grow a file slightly. The MKV came back 1,780 bytes larger as MP4 because container overhead differs. Still under the target, but worth knowing if you are chasing the last kilobyte.
What this page is not
It is not a comparison against desktop encoders, and it is not a promise about your machine. A browser can only use the encoders the browser exposes; on a phone with a hardware encoder the same job may be quicker, and on a very old laptop it may be several times slower. If your device is slow, lower the target resolution by picking a smaller cap — that reduces work far more than any setting in this page.