LosslessCut

LosslessCut trims and rearranges video and audio files without re-encoding them. It reads a file, lets segments be marked on a timeline, then copies the selected streams straight into a new container. Because the picture data is never decoded and encoded again, the operation finishes in roughly the time it takes to read and write the file, and no generation loss is introduced.

That approach carries one condition worth knowing before use: cuts land on keyframes. A start point set between keyframes moves to the nearest one, so the result can begin slightly before or after the chosen instant, and frame-exact cutting requires re-encoding the affected part. Beyond trimming, the program merges files, splits one file into several segments in a single pass, extracts individual audio, video or subtitle streams, remuxes between containers, edits track metadata and pulls still frames out of the video.

It is a desktop front-end to FFmpeg, so the formats it accepts follow what FFmpeg can demux and mux.