Turn text into subtitles
Paste a transcript, script or lyrics and get a timed SRT draft, paced by reading speed. Then align it to your audio automatically. No upload.
Sometimes the words already exist and only the timing is missing: a transcript someone typed, a translation delivered as a document, the script you are about to narrate, a page of lyrics. This tool cuts the text into cues — by line, by sentence or by paragraph — and gives each one a duration proportional to how long it takes to read, so the result is evenly paced and immediately usable as a draft. What it cannot do is know when anything is actually said, and it does not pretend otherwise: the next step is aligning the draft against real audio, which is one click away.
Worth knowing
- Every duration here is an estimate derived from reading speed, not from sound. Treat the output as a draft to align, not as a finished subtitle track — the alignment step is what makes it match a video.
- Sentence splitting knows that "Mr. Smith", "3.5 seconds" and "J. R. R. Tolkien" are not three sentences each, so a transcript written in prose does not shatter into fragments at every full stop.
- Lines are wrapped at a word boundary to your character limit, and a word longer than the limit is left intact rather than cut in half.
- When wrapping would produce more lines than allowed, the overflow is joined onto the last line instead of being dropped. A cue that is too long is a problem the validator will show you; a cue that quietly lost half its sentence is one nobody notices.
- Nothing is ever timed to overlap: each cue is followed by the gap you set, so the draft passes an overlap check before you have touched it.
FAQ
How do I make the timing actually match my video?
Generate the draft, then send it to the automatic sync tool, which listens to your video or audio file and slides the whole track until the speech and the cues line up. For fine work after that, the cue editor retimes one line at a time against local playback.
What reading speed should I use?
The default of 15 characters per second is comfortable for most viewers. Broadcast guidelines usually treat 21 as the upper limit for adult programming and slower — 12 to 15 — for children. Lower numbers produce longer cues, which is safer for a draft you will align afterwards.
Can I get VTT, ASS or something else instead of SRT?
Yes — pick any of the eleven output formats before downloading. The timing model is identical; only the container changes, so a draft destined for a web player can come straight out as WebVTT.