video tool
SRT to text
Paste an SRT or WebVTT file and the cue numbers, timings and tags come out, leaving only the words, joined into readable paragraphs. Speaker names can stay at the start of each paragraph or go, bracketed sound notes such as [door opens] are removed unless you keep them, and the repeated lines of rolling auto captions are dropped so each sentence appears once.
Loading the tool…
- Runs on your device, nothing is uploaded
- Free, no sign in
- No watermark, no file limit
How it works
- 1
Paste or open the subtitles
Paste an SRT or WebVTT file into the top box, press Open, or drop the file on the tool. The transcript appears below as soon as there is a timing line.
- 2
Choose what to keep
Tick or untick Keep speaker names, Keep [sound] and (aside) notes, and Start each paragraph with its time.
- 3
Set the paragraph pause
A new paragraph starts after a sentence ends and the next cue begins after a pause of at least this long. Two seconds suits most talks.
- 4
Copy or download
Copy puts the transcript on the clipboard. Download .txt saves it as UTF-8 text named after your file.
How the lines become paragraphs
Subtitle cues break speech wherever the screen runs out of room, often mid sentence, so stitching them back together is the real work. Each cue's text is cleaned of tags and joined to the one before with a space. A new paragraph starts only when two things are true: the text so far ends a sentence, and there is a pause before the next cue at least as long as the setting. A change of speaker always starts a new paragraph, and a very long run without a pause is broken at the next full stop after about 160 words, so a lecture does not become one wall of text. A cue that holds only a repeated line or a sound note does not count as speech, so the quiet it covers still counts towards the pause.
Speaker names
Captions mark speakers in a few ways, and the tool reads the common ones. WebVTT voice tags such as <v Priya> are the clearest. Broadcast style captions put the name in capitals before a colon, PRIYA: or >> PRIYA:. Some files use Priya: in ordinary case, which is risky to read as a name, since a line like Note: this is important would match too, so a Title Case label counts only when the same name opens two or more lines. With Keep speaker names ticked, each paragraph starts with its speaker; unticked, the labels are taken out and the words stay.
Rolling captions and sound notes
Auto generated captions on video platforms are often written as roll up lines, where every cue repeats the line before and adds one more, which is how the text appears to scroll. Read straight off the file, each sentence appears twice. Any line identical to one of the last two kept lines is skipped, which removes those repeats without touching real repetition further apart. Captions written for deaf and hard of hearing viewers also describe sounds, such as [applause] or (laughs), and music with note symbols. Those are removed by default because a transcript is usually read for the words; tick the box to keep them.
Limits worth knowing
Everything here is a guess from the text, not from listening. If the captions have no punctuation, which is common in raw auto captions, sentences cannot be found and paragraphs will only break on speaker changes or after very long runs. Square brackets and parentheses are treated as sound notes when removal is on, so a spoken aside written in parentheses goes too; untick the box if your captions use them for dialogue. Files with no timing lines at all are refused with a message, since there is nothing to strip. Nothing is uploaded and nothing is stored, ever.
Limits
- Sentence joins are guessed from punctuation. A paragraph only breaks after a full stop, question mark or exclamation mark, so captions without punctuation run together.
- A Title Case Name: counts as a speaker only when it opens two or more lines; a name used once is left in the text.
- Removing bracketed notes also removes any spoken words that the captions put in brackets or parentheses.
What it does not do
- It does not transcribe audio or video. It needs a subtitle file that already exists.
- It does not correct spelling or punctuation in the captions.
- It does not translate.
Questions people ask
Can I get a transcript from a YouTube video this way?
Yes, if you have its caption file. The owner of a video can download captions from YouTube Studio as .srt or .vtt, and auto captions come as roll up lines whose repeats this tool removes. For a video you do not own, use the transcript panel on the video page instead, and respect the creator's rights over the text.
Why is my whole transcript one long paragraph?
Paragraph breaks need a sentence ending followed by a pause. Captions without punctuation never end a sentence, and captions with no gaps between cues never pause. Shorten the pause setting to 1 second first. If the captions have no punctuation at all, breaks fall only on speaker changes and after about 160 words, so add punctuation in the source.
What do the start times look like when that option is on?
Each paragraph begins with its start time in square brackets as hours, minutes and seconds, such as [00:03:15]. That is enough to find the passage in the video again, which is useful for show notes, quotes and edits. Times come from the first cue in the paragraph, so they mark where the speaker began, not each sentence.
Can I paste the transcript into a word processor?
Yes. The output is plain text with a blank line between paragraphs, which every word processor reads as separate paragraphs on paste. There is no formatting to fight with, so headings or bold quotes are yours to add. The .txt download is UTF-8, so accented names and non Latin scripts survive when opened.
Does it count words for a reading or speaking time estimate?
It reports the word count of the transcript under the options, counting only the spoken words, not speaker names or time stamps. For an estimate of how long the text takes to read aloud at a chosen pace, paste the transcript into the speaking time calculator, which works from the same kind of plain text.
Turn the transcript into a short video
Take what you made here into videos.
Create a video with OneCraftMore tools like this
- SRT to VTT converterConvert SRT to VTT in your browser: commas in timings become dots, a WEBVTT header is added, and unsupported tags are dropped and listed. No upload.
- VTT to SRT converterConvert VTT to SRT in your browser: cues are numbered, dots in timings become commas, and positioning, regions and styles are dropped and listed. No upload.
- Shift subtitles forward or backShift subtitles earlier or later by any number of seconds, stretch them for a frame rate change, or sync two cues to the video. SRT and VTT, in your browser.
- Speaking time calculatorPaste a script and see it laid out on a clock, section by section, against the slot you have. Six measured speaking rates, and a rehearsal timer.
- Word counterCount words, characters, sentences and paragraphs as you type, and see which sentences are hard work. Runs in your browser, nothing is uploaded or stored.
Background reading: the question page on this.
Written and checked by the OneCraft team. Last checked .