How to use Video Compressor
- 1Drop in a video file — it stays on your device
- 2Choose High quality, Balanced or Smallest file
- 3Start the compression and watch the real ffmpeg progress
- 4Compare the finished size against the original
- 5Download the compressed MP4
What Video Compressor does
Video compression works by storing the differences between frames rather than each frame in full. A static shot of someone talking has almost no change from frame to frame, so it compresses enormously; a handheld shot of moving leaves changes everywhere at once and compresses badly. This is why two clips of identical length and resolution can differ tenfold in size.
Because of that, asking for a specific output size is the wrong question — the encoder cannot know in advance how compressible your footage is. What you can specify is a quality target, and let the size fall where the content dictates. That is what the three levels here do, and it is why the same setting produces a small file from an interview and a large one from a drone shot.
When to use it
- Files too large to send
- Email attachments cap out around 25 MB and messaging apps compress aggressively on their own. Compressing first means you control the quality rather than letting the platform decide.
- Phone footage that fills storage
- Modern phones record at high bitrates for editing headroom that most clips never need. Re-encoding at Balanced typically cuts the size several-fold with no difference visible on a phone screen.
- Video for a website
- The output is MP4 with H.264 and AAC, which every browser plays, and it is written so playback can begin before the download finishes rather than waiting for the whole file.
- Uploads with a size cap
- Course platforms, job applications and forms often refuse anything over a fixed size. Smallest file mode is usually enough to get under the line without making the video unwatchable.
Video Compressor settings explained
- High quality
- CRF 22. Visually very close to the source for most footage, with a moderate size reduction. The right choice when the video will be watched full-screen or edited further.
- Balanced
- CRF 27, the default. Where most people stop looking for artefacts on a normal screen, and usually a substantial size reduction. Suitable for sharing and for web playback.
- Smallest file
- CRF 32. Softening and blocking become visible on detailed footage, but the file is dramatically smaller. Use it when getting under a size limit matters more than fidelity.
- Video codec
- H.264 via libx264 at the veryfast preset. H.264 is chosen for compatibility — it plays everywhere — and veryfast because the encoder is running in your browser rather than on a server.
- Audio
- Re-encoded to AAC at 128 kbps, which is transparent for speech and close to it for music. Audio is a small fraction of a video's size, so there is little to gain from compressing it harder.
- Fast start
- The output is written with its index at the front of the file, so a browser can begin playing while the rest downloads instead of waiting for the whole thing.
What the three levels mean
| Level | CRF | Visible difference | Use when |
|---|---|---|---|
| High quality | 22 | Essentially none | Full-screen viewing, further editing |
| Balanced | 27 | Hard to spot in normal viewing | Sharing, web playback |
| Smallest file | 32 | Softening on detailed shots | Hard size limits |
Limits and what to expect
- Everything is encoded on your device, so a long or high-resolution video takes minutes rather than seconds.
- The first run downloads a ~32 MB engine, which is then cached for later use.
- Output is always MP4 with H.264 and AAC — the resolution and frame rate of the source are kept.
- Encoding is single-threaded, because the multi-threaded build requires cross-origin isolation the site does not use.
- Very large files can exhaust the browser tab's memory; a shorter clip or a lower resolution is the way around it.
Troubleshooting
- The file got bigger instead of smaller
- The source was already compressed harder than your chosen level. Re-encoding cannot recover that, and adds a generation of loss. Choose Smallest file, or accept that the video is already efficiently encoded.
- It is taking a very long time
- The encoder runs in your browser on a single thread, so a several-minute high-resolution video is genuinely slow. Trim to the part you need first, which reduces the work in direct proportion to the length removed.
- The result looks blocky
- Smallest file mode at CRF 32 shows artefacts on detailed or fast-moving footage. Move up to Balanced, and if the file must be small, reduce the resolution first — fewer pixels at higher quality usually looks better than many pixels at low quality.
- The browser ran out of memory
- Everything is held in the tab while encoding. Close other tabs, or split the video and compress the parts separately. Long 4K files are the usual cause.
Your files and your privacy
The video is processed by an ffmpeg build compiled to WebAssembly and running inside the browser tab, on your own device's CPU. No frame is uploaded, nothing is queued on a server, and no copy remains anywhere after the tab closes. The only network request is for the encoder itself, which is generic code shared by every visitor.
What to use next
FAQ
How much smaller will my video be?
It depends entirely on the footage and how it was originally encoded. A static interview at Balanced often drops several-fold; a high-motion clip already compressed efficiently may barely change. This is why the tool offers a quality target rather than promising a size — the encoder cannot know how compressible your content is until it processes it.
Does compressing a video reduce quality?
Yes, it is lossy by design. The question is whether the loss is visible. CRF 22 is essentially indistinguishable from the source for most footage, CRF 27 is hard to spot in normal viewing, and CRF 32 shows softening on detailed shots. Since your source was almost certainly already compressed, this is a second generation either way.
What format is the output?
MP4 with H.264 video and AAC audio, regardless of what you put in. That combination plays on every browser, phone, TV and editing program in current use, which is the point — a compressed file that will not open is no use.
Why is it so much slower than an online converter?
Hosted services encode on server hardware with many cores and often dedicated video chips. This runs on one thread in your browser. The trade is deliberate: it is slower, but your video never leaves the device and there is no queue, no upload wait, and no size tier to pay for.
Can I compress a video to a specific file size?
Not directly. Reaching an exact size requires either a two-pass encode with a computed bitrate or repeated attempts, both far too slow in a browser. Smallest file mode gets under most limits; if it does not, trim the clip or reduce the resolution before compressing.
Is my video uploaded anywhere?
No. Encoding happens inside your browser using WebAssembly. Nothing is transmitted or stored, and once the encoder has downloaded, the tool works with the network disconnected.