Video Compressor
Shrink a video without sending it anywhere. The tool reads the file first, reports what it is and which resolution is actually worth keeping — then shows you exactly how much smaller the result turned out.
MP4 · MOV · MKV · WebM · AVI — the larger the file, the more there is to win
The two dials that matter
Compression boils down to resolution and quantiser. Dropping 4K to 1080p removes roughly three-quarters of the pixels and is the single biggest saving available — often more than halving the file with no perceptible difference on a phone or laptop screen. Once the resolution is settled, the preset controls the quantiser: how aggressively detail is thrown away per frame. Higher values discard more high-frequency detail, which is why a heavily compressed video looks soft and blocky in motion while stills look acceptable.
Why the source matters
Compressing an already-compressed file is lossy on top of lossy, so a re-encoded 8 Mbps clip will not shrink nearly as much as a camera original that arrived at 60 Mbps. That is why this tool probes your file before it touches it: the report tells you the actual codec, resolution and bitrate, and the Auto resolution option only downscales when there are pixels worth giving up. Upscaling is never offered — it costs time and makes files bigger for no gain.
One honest caveat: re-encoding in WebAssembly runs several times slower than native FFmpeg because there is no hardware encoder to call on. A long 4K file will take a while, and the progress bar will say so. If your goal is simply to cut a clip shorter, the trimmer in stream-copy mode does that in seconds with zero quality loss — always the better first move.
FAQ
What is a good compression result?
For web delivery, 1080p at 4–6 Mbps usually looks clean and lands around 5–8 MB per minute. If you have halved the size with the Balanced preset and the result still looks right to you, stop there.
Should I keep or re-encode the audio?
Keep it. Audio is a tiny fraction of a video's size and re-encoding only adds loss and time. Only re-encode when the source audio is in a codec the target container cannot carry.
Is the output watermarked or trimmed in length?
No. The result is your file, re-encoded — full length, full frame, no watermark, and nothing about it was sent anywhere.