When you have text but no timings
Subtitle files need a start and end time for every line. If you already have the words — a script you wrote, a transcript exported as plain text, meeting notes you want as captions — the missing piece is timing, not text.
Without audio there is nothing to measure, so this tool estimates: each line is given time in proportion to how many characters a viewer has to read. That is not a guess pulled from nowhere — it is the same calculation a subtitler makes when they read a line aloud and decide it needs two seconds.
When this is the right tool, and when it is not
- Right: a script you are about to record. The estimated timings are usually close enough to drop in and nudge.
- Right: a transcript from somewhere else that lost its timings, where approximate sync is acceptable.
- Wrong: a video that already exists. Transcribing it gives real timings measured from the audio, which will always beat an estimate.
- Wrong: anything where the sync has to be frame-accurate — musical cues, lip-sync, comedy timing. Those need a subtitle editor.
How the estimate works
Three rules, applied in order:
- Split the text into sentences. Each sentence becomes one or more cues.
- Wrap any sentence longer than the line limit, preferring to break at a comma or semicolon rather than mid-phrase.
- Give each line
characters ÷ reading speedseconds, with a floor of one second and a ceiling of seven. Short fragments get merged with their neighbour rather than flashing briefly.
The defaults — 42 characters, 15 characters per second — come from broadcast subtitle practice. Lower the reading speed if the audience is reading in a second language or the video is dense.
After you convert
The output is standard SRT, which every editor and platform accepts. If your player wants VTT instead, the difference is the header line and a period before the milliseconds — see the format comparison. To convert an existing file rather than generate one, use the VTT to SRT converter (it handles both directions).