Free tool
YouTube Transcript Generator
Paste a link and get the video's transcript — as readable paragraphs, as timestamped lines, or as a subtitle file. No sign-up, no watching required.
- 1 Paste a linkAny video, Shorts or youtu.be link
- 2 Read itParagraphs, timestamps or raw captions
- 3 Take it with youCopy, or download .txt, .srt or .vtt
How do you get a transcript from a YouTube video?
Paste the link above. The tool reads the captions the video already has and lays them out as paragraphs you can actually read — auto-generated captions arrive as hundreds of two-second fragments, which is why copying them out of YouTube gives you a wall of broken lines. No words are changed: only where each line starts and stops.
How to use the transcript generator
- Paste any YouTube link. A watch link, a youtu.be short link or a Shorts link all work. Nothing to install and no account.
- Pick the shape you need. Paragraphs for reading and quoting, timestamps for following along with the video, raw captions for editing subtitles line by line.
- Copy it, or download it. Plain text for writing, or a
.srtor.vttsubtitle file with the original timings intact.
Why a copied transcript looks like that
YouTube's own transcript panel shows captions the way the player needs them: one short fragment every couple of seconds, each stamped with a time. Select it, copy it, and you get exactly that — several hundred lines, most of them half a sentence, with a timestamp wedged in front.
That format exists so words can appear on screen in time with speech. It was never meant to be read. Turning it into prose means joining the fragments back up and deciding where a paragraph ends, which is what this tool does.
Where the paragraph breaks come from
Not from guesswork about grammar. Every caption carries a start time and a length, so the tool can see where the speaker actually stopped talking. A paragraph ends at a real pause, or at a full stop, or once it has simply run long enough to read comfortably — whichever comes first.
What it will not do is invent punctuation. Auto-generated captions often have none at all, and a tool that adds commas and capital letters is writing words the speaker did not say. If you want that, the summarizer rewrites the content openly and says so. Here, the words are the speaker's and only the line breaks are ours.
One small thing the rules do not cover: a leftover line or two at the very end is joined to the paragraph above rather than left standing alone. “All right, thanks for watching” is not a paragraph.
The three views, and what each is for
| View | Use it when |
|---|---|
| Paragraphs | You're reading, quoting, or pasting the content somewhere else |
| Timestamps | You want to find a moment and jump to it — every timestamp is a link |
| Raw captions | You're editing subtitles and need the original lines and timings |
Switching views costs nothing and fetches nothing: the captions are already in your browser, so the three formats are three ways of drawing the same data.
How good the captions are depends on who wrote them
Every transcript here is exactly as good as the captions behind it, and there are two kinds. Captions the uploader wrote or corrected are usually close to perfect, with punctuation and proper spelling of names. Captions YouTube generated are a speech-recognition guess: no punctuation, no speaker labels, and reliable mistakes around names, jargon, accents and anything said over music.
| Caption source | What it reads like |
|---|---|
| Written by the uploader | Punctuated, correctly spelled, usually quotable as-is |
| Auto-generated | Accurate on ordinary speech, wrong on names and technical terms, no punctuation |
| Auto-translated | Not used here — you get the captions in the language they were published in |
Nothing in this tool corrects them, which is the point: if you are going to quote someone, you want what the captions actually say, not a tidied version that might have tidied away the thing you were checking. Always confirm a direct quote against the audio at the timestamp before you publish it.
Reading a two-hour video in five minutes
Fetch the transcript, switch to the timestamps view, and use your browser's own find (Ctrl-F, or Cmd-F on a Mac). Search for the term you came for, read the surrounding lines, then click the timestamp to open the video at that second and check the claim with your own ears. That loop is far faster than scrubbing a progress bar, and it is the reason the timestamps are links rather than decoration.
Subtitle files, with the timings kept
The .srt and .vtt downloads carry the original start
and end times, so they load straight into a video editor, a player, or another
platform's subtitle field. Where captions overlap — which happens with
auto-generated ones — each cue is trimmed to end when the next begins, because
a file with overlapping cues is rejected by some players and silently mangled
by others.
Transcript, captions, subtitles
The three words get used interchangeably and mean slightly different things.
Captions are the timed text YouTube shows over a video.
Subtitles are the same thing in a file a player can load, which
is what .srt and .vtt are. A
transcript is all of it as continuous text, without the timing
— the thing you read rather than watch. This tool starts from the first and
gives you any of the three.
Videos this can't do
It reads captions that exist. It does not run speech recognition, so a video nobody has captioned and that YouTube hasn't auto-captioned has nothing to read, and you'll be told that rather than handed a plausible invention.
| What you'll see | What it means |
|---|---|
| No captions | The uploader disabled them and auto-captions didn't run — common on music and very new uploads |
| Restricted | Age-restricted, members-only or region-locked, so the captions can't be read either |
| Private or removed | Nothing to fetch. Unlisted videos do work if you have the link |
What to do with a transcript once you have it
- Quote it accurately. The words are the speaker's, with the time you can point to.
- Search a long video. Press Ctrl-F on the transcript instead of scrubbing a two-hour podcast.
- Build chapters. The timestamp view shows where each part starts; the description generator turns that into chapters YouTube will accept.
- Tidy it further. If you need [Music] tags and stray markers stripped, the transcript cleaner does that to any transcript you paste in.
- Get the argument, not the words. Summarize the video and you get the takeaways and a chapter map instead of 9,000 words.
Transcript or summary?
A transcript is everything that was said, which is what you want when the exact wording matters — a quote, a claim, a recipe step, a legal or medical detail. It is also long: an hour of speech is around 9,000 words, roughly a 40-minute read. A summary is the argument in a couple of minutes, and it is the right tool when you're deciding whether the video is worth your time at all.
Frequently asked questions
Is it free?
Yes, with no account. There's a daily cap per connection so the tool stays available to everyone, and transcripts already fetched are served from cache without counting against it.
Can I get a transcript of a video with no captions?
No. This reads captions that already exist rather than transcribing audio, so a video with none has nothing to read. You'll get a clear message instead of a guess.
Does it work for other languages?
Yes — you get the captions in whatever language they were published in, and the language is shown above the transcript. It doesn't translate.
How accurate is it?
Exactly as accurate as the captions. Captions written by the uploader are usually near-perfect; auto-generated ones mishear names, technical terms and anything said over music. Nothing here corrects them, so what you read is what YouTube has.
Can I get it without the timestamps?
That's the paragraphs view, and it's what you get by default. Copy or download from there and no timestamps come with it.
Does it work on unlisted or private videos?
Unlisted videos work if you have the link. Private ones don't — nobody without access can read them, this tool included.
Can it translate the transcript?
No. You get the captions in the language they were published in. Translating would mean handing you words nobody said, which is the one thing this tool avoids.
Can I use the transcript in my own work?
Quoting with attribution is normal practice. Republishing someone's full transcript as your own content isn't — the words belong to whoever said them, and the same copyright applies as to the video.
Does it work on long videos?
Yes. A three-hour podcast takes longer to fetch, and the tool keeps you posted while it works rather than timing out.
Nine thousand words is a lot to read
Get the argument, the takeaways and a chapter map instead — from the same video.
Summarize a video — free