How to use Video Trimmer
- 1Drop in a video file
- 2Enter the start and end times in seconds
- 3Run the trim — the streams are copied, not re-encoded
- 4It finishes in seconds even on a large file
- 5Download the trimmed clip
What Video Trimmer does
Trimming here does not re-encode anything. The video and audio streams are copied out of the source and into a new container exactly as they are, which is why it finishes in seconds on a file that would take minutes to compress, and why the result is pixel-identical to the original.
That speed has one consequence worth understanding. Compressed video stores full frames only occasionally — keyframes — and describes everything between them as differences. A copy cannot start midway through that chain, so the cut lands on the nearest keyframe rather than the exact second you asked for. Depending on how the video was encoded, that can be anywhere from a fraction of a second to several seconds out.
When to use it
- Cutting a clip out of a long recording
- Pulling two minutes from an hour-long screen recording or stream, without waiting for the whole thing to re-encode.
- Removing dead time at the start
- Recordings usually begin before anything happens. Trimming the front is the single most common video edit there is.
- Getting under a length limit
- Platforms cap upload duration by tier. Trimming to fit costs no quality, unlike compressing to fit.
- Quoting a moment
- A short section for a message or a report, taken from the source without degrading it, so the detail being pointed at is still visible.
Video Trimmer settings explained
- Start time
- In seconds from the beginning. The trim actually starts at the nearest keyframe at or before this point, so the clip may include a little more than you asked for.
- End time
- In seconds from the beginning, not a duration. It must be greater than the start time, and a value that is not is rejected rather than silently swapped.
- Stream copy
- Always used, and not optional. Both streams are copied rather than re-encoded, which is what makes the operation fast and lossless — and what causes the keyframe behaviour.
- Output container
- The source container is kept where it can hold the copied streams — MP4, WebM, MOV, MKV or AVI. Anything else is written to MKV, which accepts almost any codec combination.
Trimming compared with the other video operations
| Operation | Re-encodes? | Speed | Quality |
|---|---|---|---|
| Trim | No | Seconds | Identical to source |
| Mute | No | Seconds | Identical to source |
| Compress | Yes | Minutes | Reduced by design |
| Resize | Yes | Minutes | Reduced by design |
Limits and what to expect
- Cuts land on keyframes, so the start can be up to several seconds earlier than requested.
- Only one section per run — removing a middle portion means trimming twice and joining the results elsewhere.
- The output stays in a container that accepts the copied codecs, which may not be MP4.
- Files whose container cannot hold the copied streams are written to MKV, which some players handle poorly.
- Very large files are still bounded by the browser tab's memory, even though no encoding happens.
Troubleshooting
- The clip starts earlier than I asked for
- That is the keyframe interval. A stream copy has to begin at a keyframe, so it uses the nearest one at or before your start time. Videos recorded by screen capture software often have keyframes several seconds apart, which makes this more noticeable.
- I need the cut to be frame-accurate
- That requires re-encoding, which trades the speed and the losslessness for precision. Trim generously here, then compress the result — the compression pass can start anywhere.
- The output is an MKV, not an MP4
- The source codecs could not be copied into the original container, so MKV was used because it accepts nearly anything. Convert it if you need MP4 specifically — that pass does re-encode.
- The end time was rejected
- Both fields are positions in seconds from the start of the video, not a start and a duration. An end of 30 with a start of 45 is rejected because it would describe a negative-length clip.
Your files and your privacy
The trim runs inside your browser through an ffmpeg build compiled to WebAssembly. The video is never uploaded and no copy is written to a server. Because a stream copy does no encoding work, this is also the one video operation where processing locally costs you almost nothing in time compared with a hosted service.
What to use next
FAQ
Does trimming a video reduce its quality?
No. The streams are copied rather than re-encoded, so the trimmed clip is bit-identical to the corresponding section of the source. This is the only lossless way to shorten a video, and it is why the operation takes seconds rather than minutes.
Why did my clip start before the time I entered?
Compressed video only stores complete frames at keyframes, and a stream copy has to begin at one. The trim uses the nearest keyframe at or before your start time, so the clip can begin up to several seconds early depending on how the source was encoded.
How do I get a frame-accurate cut?
You cannot with a stream copy — precision and losslessness are mutually exclusive here. Trim slightly wide, then run the result through the compressor, which re-encodes and can therefore start on any frame.
Can I remove a section from the middle?
Not in one pass. Trim the part before it and the part after it as two clips, then join them in a video editor. This tool produces one continuous section per run.
Why is trimming so much faster than compressing?
Because it does no encoding at all. Compression decodes every frame and re-encodes it, which is expensive. Trimming copies the existing compressed data and only rewrites the container around it.
Is my video uploaded?
No. Everything runs in your browser through WebAssembly, on your own device. Nothing is transmitted or stored.