MMP: Minimalist Media Player

The actual editing of your video or audio (cutting into segments and stitching them together) is performed by FFmpeg.

FFmpeg is capable of some very complex audio and video manipulation, which is why it has become almost the de facto standard for the majority of audio/video editing and conversion tools available today¹⁾.

However, this complexity comes at a cost:
The more complex the operation, the more complex the parameters are that you have to supply to FFmpeg and the more chances that FFmpeg will complain about conflicting parameters, or parameters that conflict with the contents of your input file.

For this reason, MMP keeps it [relatively ²⁾] straight forward to [all but] guarantee successful editing.

Conversely though, this also comes at a price, which is that cutting is performed at the “key frame” level, not at user-specified frames.

For most general editing this is entirely acceptable. ³⁾

The great advantage that this approach does have, though, is that producing the edited output file from your required edits is blisteringly fast - a couple of seconds and it's all over. ⁴⁾
Most editors re-encode the segments to a brand new video, which can take a considerable amount of time;
MMP simply does a copy/paste: it extracts parts of the original file and stitches them together. This also means the video quality is exactly that of the original file. ⁵⁾

(a layman's description)

Video [file] compression works by only storing what has changed from frame to frame.
Every second or so, depending on how much the video image is actually changing, a Key Frame will be written ⁶⁾ - a complete-image frame to re-establish a “difference” starting point.
The majority of a video file these days is just the changes between Key Frames.
This greatly reduces the size of video files.

It is those full-image Key Frames that MMP and FFmpeg use to make your cuts.

In a nutshell, if you make a cut (start a segment) between key frames ⁷⁾, FFmpeg will make the actual cut on the most recent key frame preceding your cut.
Similarly, the cut at the end of a segment will also be made at the key frame before your cut.

In practice, then, we can be reasonably confident that:

• the start of a segment is likely to include the frames you require
• the end of a segment probably needs to be a second or two longer than you require, so that it includes the next key frame on which FFmpeg can make the actual cut.

This is also why

1. it is very important to check your edited file before deleting the original ⁸⁾
2. having two adjacent segments included in your edited video will likely result in overlap* ⁹⁾