Atak,
I don't recall if I've ever asked this before, although I get the feeling I may have broached the subject in years previous. I'm curious about what method you are employing to split the files (I don't see anything "obvious" in the tools directory except an executable called vjoin, which I assume is for joining). I'm trying to do some research on the best way to accomplish splitting/joining accurately and efficiently without any glitches. I've seen references online to using ffmpeg to split on keyframes (and ffprobe to acquire an index of keyframes to help with this) but haven't really found any real turnkey solution, outside of something like DVE
https://github.com/nergdron/dve
slightly related, do you have any plans to ever expose ffmpeg for use by the user (as opposed to x264/x265)? I'm mainly interested in targeting and accelerating VP9 (libvpx for now) across multiple cores/machines, but down the road when Intel's SVT libraries become more mature this could be useful to a lot of people as well, especially with AV1