Compress Video
Choose one video up to 50 MB, pick a quality and a resolution cap, and get a smaller MP4 you can play before you save it. The whole job runs inside your browser, so the video never leaves your device.
How to use
- Choose one video: Drop a video onto the box or click to pick it. One file per run, up to 50 MB, and only video files are accepted — a 50 MB-plus file or an audio file is skipped with a note saying so.
- Pick a quality: High (larger file) keeps the most detail, Recommended (balanced) is the default, and Smaller (size first) produces the smallest result. Start with Recommended and only step down if the file still doesn't fit.
- Pick a resolution cap: Leave Resolution on Up to 1080p, choose Up to 720p for a smaller picture, or Keep original resolution. A video already below the cap is never enlarged.
- Compress, check, save: Click Compress video. The first run spends a while fetching the engine before any percentage appears, and you can press Cancel at any point. When it finishes you get the size before and after, a player to check the result, and Download (Save to device on a phone).
How to compress video in your browser
Compress video one file at a time, up to 50 MB, in your browser. Choose a quality level and resolution cap to create an H.264 MP4 without uploading the file. The result may be smaller, unchanged or larger, depending on the source and settings.
The video is re-encoded on your own computer or phone by ffmpeg.wasm, a build of FFmpeg that runs inside the browser tab. Nothing is uploaded: there is no server fallback in the code at all, so a file that your device can't handle fails with a message rather than being sent somewhere to be processed. Close the tab and nothing is left behind.
Before encoding starts, the tool runs one quick pass to read the video's length, codec, resolution, rotation and shoot date. If that pass can't make sense of the file — a damaged download, a truncated recording, or something that only has a video extension — it stops there and tells you the file isn't a video it can read, instead of grinding through a doomed encode.
The result is always an MP4 with H.264 video, whatever you put in. MOV from an iPhone, MKV, WebM and AVI all come out as .mp4, because this is a real re-encode rather than a change of container. The playback index is written to the front of the file, so the MP4 can start playing while it is still downloading — useful when you drop it in a shared drive or attach it to a message.
Because re-encoding needs the multi-threaded engine, the page has to run with cross-origin isolation and your browser has to support SharedArrayBuffer. An up-to-date Chrome, Edge or Firefox does; private windows and some in-app browsers switch the capability off, and there you are stopped before the download starts rather than after it.
Which quality and resolution should you choose?
High → CRF 23 → prioritize detail Recommended → CRF 28 → starting point Smaller → CRF 32 → prioritize file size 1080p cap → short edge up to 1080 px 720p cap → short edge up to 720 px
Quality is set as a constant rate factor (CRF), the number H.264 uses to decide how much data each scene deserves: a busy, moving shot gets more, a static one gets less. A higher CRF means a smaller file and a softer picture, so the three options run in a predictable order: on the same footage, Smaller comes out below Recommended, and Recommended below High.
The resolution cap applies to the short edge of the picture: Up to 1080p limits a landscape clip to 1080 pixels tall and a portrait one to 1080 wide, and Up to 720p does the same at 720. It only ever shrinks. In our reference run a 160×120 clip stayed 160×120 under both Up to 1080p and Up to 720p, and the file size was identical — if your footage is already small, the resolution setting does nothing and the quality setting is the only lever.
A good default: Recommended with Up to 1080p for anything you want to look right on a laptop or TV, and Smaller with Up to 720p when a size limit matters more than detail. For 4K footage, try the 1080p cap first and inspect fine detail in the preview before saving.
Why some videos barely shrink
Compression has nothing to recover from a video whose bitrate is already low, and such a file can even come out slightly larger. Two mechanisms explain it. First, the quality setting is a target, not a reduction: if the original already used less data than the CRF you picked would spend, the encoder may well spend more than the original did rather than less. Second, the sound is usually a meaningful share of a short clip, and when the first audio track is already AAC the sound is copied across untouched to avoid a second round of loss — so that part of the file cannot shrink at all.
As a worked example, a 26,714-byte test clip (160×120, 2.0 seconds, H.264 video at roughly 26 kbps plus mono AAC audio at 69 kbps) was run through the exact parameters this tool uses, using desktop FFmpeg rather than the browser: High came out at 26,794 bytes — 0.3% bigger — while Recommended landed at 25,372 bytes (5.0% smaller) and Smaller at 24,522 bytes (8.2% smaller). Those are reference figures for one tiny low-bitrate clip, not a promise for your file; we have no measured compression ratio for a typical phone recording, so this page doesn't quote one.
The tool is honest about the outcome instead of hiding it. When the result barely moved it says so and points you towards a lower quality or a lower resolution cap, and when the result came out bigger it says that too — and if you were already on Smaller with Up to 720p, it tells you to keep the original instead of chasing it further. Both lines send you back through Start over, which clears the result so you can choose the original file again with different options.
What changes in the compressed video
All of the global metadata is deliberately thrown away: GPS coordinates, device make and model, and the name of the software that produced the file. That is the privacy-relevant part, and it is the main reason to run a holiday clip through this tool before posting it. The one thing written back on purpose is the date and time the video was shot, so the file still sorts correctly in a photo library.
The picture of the first video track and every audio track survive. The sound is decided once, by the first audio track: if that track is AAC, every audio track is copied across untouched, and if it is anything else, every track is re-encoded to AAC at 128 kbps. For the single soundtrack that nearly every recording has, that means AAC is passed through without a second round of loss. A file whose first track is AAC and whose second track is something else keeps that second track in its original codec, so the output is not guaranteed to be AAC throughout. A video with no sound stays silent. Orientation is kept, so a portrait clip stays portrait. Track-level language and title tags and chapters are not affected.
What does not survive: subtitle tracks and data tracks are dropped, as is any video track beyond the first — a cover-image track, for example. Output uses 8-bit YUV 4:2:0. This tool does not perform HDR-to-SDR tone mapping; HDR footage may display incorrectly. Keep the original of anything graded in HDR. And the compression is lossy by nature; there is no lossless mode, no option to keep the original codec or container, and no H.265 output.
There is no password or encryption concept in video the way there is in PDF, so there is nothing to unlock. A DRM-protected or corrupted file simply shows up as one the tool can't read.
Shrinking photos instead, or stripping their GPS data? Use Compress Image.
Getting a video under an email or upload limit
Start from the limit you actually face. A personal Gmail account refuses attachments over 25 MB and offers a Drive link instead, chat apps such as WhatsApp apply their own limits to video sent in a chat, and job and university portals often cap uploads at 5 or 10 MB. Since this tool accepts files up to 50 MB, the realistic job is a clip of a few dozen megabytes coming down under a cap, not a half-hour recording.
If Recommended at Up to 1080p isn't enough, press Start over and run the original again with Smaller and Up to 720p rather than compressing the output a second time — each pass throws away detail, and compressing an already compressed file mostly adds blur. If even that misses the limit, the video is simply too long for the target: trim it in your phone's own editor first, or upload it to a drive and send the link.
A few practical notes on the wait. The first run downloads about 31 MB of engine before anything can happen, which is unrelated to how big your video is; the download starts the moment you choose a file, so it overlaps with picking your options. There is no total time limit — a long job that is still making progress is left alone — but if the encoder goes 75 seconds without a single sign of life, the tool treats it as stuck and stops it. If a first attempt fails, a second one runs automatically with faster, simpler encoder settings. When everything fails you get a message saying your browser couldn't finish, usually for lack of memory, with nothing uploaded; closing other tabs, using a shorter clip or dropping to Up to 720p is the fix.
Frequently asked questions
- How much smaller will my video get?
- There is no fixed ratio, and we don't quote one, because we have no measured result for a typical phone recording. In a desktop FFmpeg reference run using this tool's settings, a 26,714-byte, 160×120, 2-second clip became 25,372 bytes at Recommended: 5.0% smaller. This is a reference example, not a browser measurement or a prediction for your video. The tool shows the size before and after, so you can judge and retry.
- Is my video uploaded to a server?
- No. The compression runs inside your browser tab with ffmpeg.wasm, and the code has no server fallback — if your device can't finish the job, it reports an error rather than sending the file anywhere. That makes it suitable for footage you wouldn't hand to a stranger, and closing the tab leaves nothing behind.
- Why does nothing happen for a while after I click Compress video?
- The first run has to fetch about 31 MB of video engine before it can start, and the percentage only appears once actual encoding begins. The download starts as soon as you choose a file, so some of that wait overlaps with picking your quality. The engine size is the same whether your video is 2 MB or 50 MB.
- Can I compress MOV or MKV to MP4?
- Yes. MOV, MKV, WebM and AVI are all accepted, and what comes back is one MP4 with H.264 video, named after your original with -compressed added. The container is not simply swapped — the picture is genuinely re-encoded, so the original codec is not kept. The result plays right in the page so you can check it before saving, and its playback index sits at the front of the file, which lets most players start it while it is still downloading.
- Does it remove the GPS location from my video?
- Yes. Global metadata — GPS coordinates, device make and model, software name — is stripped from the output. The date and time the video was shot is the one item written back on purpose, so the compressed file still sorts by date in a photo library.
- Can I compress a video to an exact size, like 25 MB?
- Not with this tool. It offers three quality steps and a resolution cap, not a size target, so getting under a limit is a matter of choosing Smaller and Up to 720p and checking the resulting size. Trimming the clip before compressing helps more than any setting when the gap is large.
- Can I compress several videos at once, or a file bigger than 50 MB?
- No — one video per run, and up to 50 MB. Anything larger is skipped with a note. The 50 MB figure is a site-wide limit; beyond it the real ceiling is your device's memory, since the whole file is handled inside the tab. For a batch, run them one after another; for a bigger file, trim it or export a smaller version from the app that made it.
- Will subtitles, HDR or the original format be kept?
- No to all three. Subtitle and data tracks are dropped, and the output is always H.264 in an MP4 container whatever you put in. Output uses 8-bit YUV 4:2:0. This tool does not perform HDR-to-SDR tone mapping; HDR footage may display incorrectly. If any of those matter for your footage, keep the original file alongside the compressed copy.