Split subtitles by time
Split an SRT/VTT/ASS/SSA/SBV file into two at a given timestamp, e.g. for a video cut into two parts. Cues are assigned to a part by their start time; the second part's timestamps are re-based to start at 0. Runs on your device.
Runs on your device. The file is never uploaded.
Split subtitles by time cuts one file into two, named with -part1 and -part2 before the extension. A cue belongs to a part by its start time, so a line spanning the cut stays whole in part one, and part two has atMs taken off every time. A split leaving either part empty is refused.
Questions
Where exactly does it cut?
At the millisecond you give in the split point setting, and a cue is assigned by its start time. A cue that begins before the split but ends after it stays whole in part one, so a line of dialogue is never cut in half. There is no default split point, because there is no sensible one to guess.
Are the second part's timings rebased?
Yes. Every cue in part two has the split point subtracted from its start and end, so the part starts at zero and lines up with a video file that also starts at the cut. Part one keeps its original times.
What are the two files called?
Your file name with -part1 and -part2 before the extension, keeping the input format: movie.srt gives movie-part1.srt and movie-part2.srt. They appear as two separate downloads, not a zip. Both keep the format that went in, so splitting a .vtt gives you two .vtt files, written by the same serializer the format converters use.
Why did it refuse my split point?
Because one part would have been empty. Splitting at or before the first cue is refused with a message naming the first cue's start time, and splitting at or after the last cue's start is refused the same way. A value that is not a non-negative number is refused with "atMs must be a non-negative number of milliseconds".
Which formats can I split?
SRT, VTT, ASS, SSA and SBV, decided by the file extension. An extension outside that list is refused with a message that lists them. The output keeps whatever format went in, and the same parser runs for both parts, so timings are re-formatted exactly rather than string-edited.