Video Trimmer

Drop the file onto the page, drag the two handles to the span you want to keep, and press Trim. The default method copies the streams without re-encoding, so the clip comes out at the original quality and is usually ready within seconds. You can also type the two times instead of dragging.

Not in Copy mode. The selected packets are moved into a new container without ever being decoded, so the output holds the same picture data the source did — there is no second generation of compression. Re-encode mode does decode and rebuild the video, which is why it is the second option rather than the default.

Because a lossless copy has to begin on a keyframe. Much of a video records changes relative to earlier frames, so starting the copy elsewhere can leave its first packets referring to frames that are not in the new file. This tool reads the keyframe positions out of your file, moves the cut back to the last one before your start point, and shows you both times. Switch to Re-encode if the cut has to land on the time you set.

No. The engine is FFmpeg compiled to WebAssembly and it runs inside your browser tab. The file is read from disk into the tab's memory and the result is written back out as a download. Nothing about the file crosses the network, which you can confirm by watching the network panel while a trim runs.

A keyframe gives a decoder a self-contained place to start. Many of the frames between keyframes record changes relative to earlier frames, which is what keeps video files small. A stream copy moves packets into a new file without rebuilding them, so beginning between keyframes could leave those first packets without the reference frames they need. Starting at a keyframe keeps the needed run of packets together. Recordings often place one every one to ten seconds; the ticks under the timeline show the positions in your file.

Removing the start is the same as keeping everything after a point: drag the left handle to where the real content begins, or scrub to it and press Start here. Removing the end is the mirror image with the right handle. Both are single continuous spans, which is exactly what a lossless copy handles best.

Copy, unless you have a reason not to. It preserves quality exactly, finishes in seconds, and keeps the original container. Re-encode earns its cost in two cases: the cut must land on a particular moment, or the keyframe before your start pulls in seconds of footage you cannot include. Re-encoding is slower on long or high-resolution clips because it runs on a single thread here.

A copy is close to the time it takes to write the file, because no frame is ever decoded — usually seconds even for a long selection. A re-encode has to decode and compress every frame, so it scales with resolution and length; a short 720p clip is quick, while a few minutes of 4K will keep you waiting.

Not here. Joining two spans means either re-encoding both or splicing at keyframe boundaries, and the failure modes are unpleasant enough that a trimmer should not pretend to handle it in passing. This tool keeps one continuous span, which covers trimming the ends and lifting a clip out of a longer recording.

It reads MP4, MOV, WebM, MKV, AVI, M4V and most other things FFmpeg understands. A copy keeps the source container, so an MP4 stays an MP4 and a WebM stays a WebM. A re-encode always writes MP4, because H.264 video does not belong in a WebM file. Files with no audio track are handled as they are.