Trim & Cut Video

1. 1

   Load your video

   Drop in an MP4, MOV, WebM, MKV or GIF. It is read on your device — nothing is uploaded.

2. 2

   Set the range

   Drag the in and out handles on the timeline, or type an exact timecode. Add more ranges to keep or remove several parts.

3. 3

   Pick how to cut

   Lossless copies the original video data and starts at the nearest keyframe. Frame-accurate re-encodes so the cut lands exactly where you put it.

4. 4

   Choose the output

   Export one joined clip or each range as its own file, in MP4, MOV or WebM, with or without the audio.

5. 5

   Download

   Save the result. Nothing is written to your disk until you press download.

Most recordings contain a fraction of what you actually want to keep. A talk has three minutes of setup before it starts, a screen capture has a false start, a phone clip has ten seconds of pocket at the end. Cutting is how a recording becomes something worth sending — and it is the one edit almost every video needs before anything else happens to it.

What cutting gets you

• Only the part that matters: pull one moment out of an hour of footage.
• Nothing to explain: a clip that starts where the point starts needs no preamble.
• Smaller files: a shorter clip uploads and sends faster, without touching the quality.
• Several cuts in one pass: remove the mistakes and keep everything else, in one export.
• Fits the platform: trim to the duration a story, short or attachment will accept.

Almost every online cutter decodes your video, re-encodes it, and hands back a new file. That is a generational loss: the picture is compressed a second time, on top of whatever compression it already carried, and it never comes back. It is also slow, because re-encoding means every single frame passes through a decoder and then an encoder.

What this tool does instead

• It copies, not re-encodes: in lossless mode the original compressed video data is moved into a new file untouched. The picture in the result is the same data as the picture in the source.
• It is close to instant: measured on a 1080p source, a four-second lossless cut completes in about 13 milliseconds, against roughly 1.7 seconds to re-encode the same range.
• It starts at a keyframe: compressed video is only decodable from certain frames, so a lossless in-point moves back to the nearest one. The timeline shows exactly where it will land before you export.
• Frame-accurate is still there: when the cut has to be exactly where you put it, the tool re-encodes only that range — and says so.

When lossless is not available

A copy is only possible when the container you are exporting to can carry the video data your source already holds. An H.264 recording copies straight into MP4 or MOV, and a VP9 recording copies into WebM — but H.264 cannot go into WebM, so choosing WebM for an H.264 source falls back to a re-encode. The tool works this out from the file itself and tells you which mode you are about to get, rather than discovering it half way through.

A single start and end point is the easy case. The harder and more common one is a recording with three good stretches and two bad ones — and cutting that with a single-range tool means five exports and a separate merge.

• Keep mode: each range you add is a piece you want. Everything outside them is discarded.
• Remove mode: each range you add is a piece you do not want. The tool exports everything else, joined.
• One clip or several files: export the kept ranges joined end to end, or as separate files delivered together.
• Overlaps are resolved: two ranges covering the same seconds are merged rather than exported twice.

Four decisions, and only the first two change what the result looks like.

• Cut method: lossless unless the boundary has to be exact to the frame. There is no quality reason to re-encode.
• Output format: MP4 plays everywhere. MOV suits Apple workflows. WebM is the one to pick when the source is already WebM.
• Audio: keep it, or drop the track entirely to save space when the sound is not part of what you are sending.
• Quality: only used when re-encoding, and set as a share of what the source already spends — so a cut can never come out with more data in it than went in.

Worth knowing

• A lossless cut cannot reduce quality, because no pixel is decoded or re-encoded at any point.
• The lossless in-point moves earlier, never later — you always get at least the range you asked for.
• Keyframes are typically one to ten seconds apart, so the snap is usually under a second.
• Re-encoding is the only way to land a cut exactly between two keyframes.
• Audio can start up to about one audio frame — roughly 20 milliseconds — after the video on a lossless cut, because audio packets do not align with video keyframes.
• Export time on the lossless path scales with the length of the clip you keep, not the length of the source.

• Social clips: pull a thirty-second moment out of a long recording for a short or a reel.
• Meetings: extract the one agenda item somebody actually needs to watch.
• Bug reports: isolate the reproduction and drop the two minutes of setup around it.
• Lectures and talks: remove the pauses, questions and false starts in one pass.
• Attachments: shorten a clip until it fits a size limit, without re-compressing the picture.

Video is not stored as a sequence of independent pictures. Most frames describe only what changed since the frame before them, and only occasional keyframes can be decoded on their own. That single fact is what shapes everything this tool does.

• Reading the file: the container is parsed on your device to find the video and audio tracks, their duration, and the exact position of every keyframe.
• Planning the ranges: your ranges are ordered, clamped to the clip, and merged where they overlap, so what is exported is exactly what is selected.
• Copying or re-encoding: in lossless mode the compressed data for the selected range is copied into the new file and the timestamps are shifted to start at zero. In frame-accurate mode the range is decoded and encoded again.
• Joining: when several ranges go into one file, each one continues where the last ended, so the result plays as a single clip.
• Checking the result: the file that comes out is read back before it is offered, and must contain a real video track with a real duration — a cut that produced nothing is reported, not handed over.

Cutting runs entirely on your device, using our own hardware-accelerated pipeline.

• Video compressor
• Video merger
• Video cropper
• Video to GIF
• Extract audio
• Resize frame

  Match resolution to destination before export