How Cuekind processes subtitles
Supported files
Use UTF-8 SRT or WebVTT. Limits: 20 files, 2 MiB per file, 10 MiB per batch and 20,000 cues per file or merged output. Timestamps use millisecond precision and stay between 00:00:00.000 and 99:59:59.999.
Conversion
Cuekind parses each cue into a start, end and visible text. Output is regenerated with sequential numbers and standardized timestamps. Styling, voice labels, cue positioning, STYLE/REGION/NOTE blocks and other metadata are not preserved. SRT and VTT output escape text markup characters. TXT output contains only cue text, with no timing.
Moving the timeline
An offset is added to both cue endpoints, rounded to the nearest millisecond. Negative results block export by default. If you explicitly choose cropping, cues ending at or before zero are removed and crossing cues start at zero. A result with no remaining cues cannot be exported.
Merging and splitting
Merging takes at least two valid input files and stably orders cues by start time, then end time. It retains duplicate or overlapping cues. Splitting clips crossing cues into both parts; timestamps in part two are rebased by subtracting the split point. Both parts must contain at least one cue.
Review flags
The checker highlights decreasing start times, overlap with an earlier cue, more than two text lines and more than 20 non-whitespace Unicode code points per second. These are review signals. Reading-speed expectations depend on language, audience and publisher.
Example
A cue from 00:00:01,000 to 00:00:04,000 shifted by +1.5 seconds becomes 00:00:02,500 to 00:00:05,500. Splitting that original cue at 2 seconds gives 1–2 seconds in part one and 0–2 seconds in part two.