§ Paste a video, read it as text
Every word, with its timecode.
Paste a YouTube link and get the video’s whole transcript, with a timecode on every line — lifted from its caption track when it has one, listened to when it has none. Free, no sign-up, five videos a day. Copy it, download it, or mark any line and take a vertical clip of it.
§ What you get
Watch links, youtu.be short links, Shorts, live and embed links all work, on the desktop and the mobile host. A channel, a playlist or a YouTube Music link is not one video, and you are told so rather than kept waiting.
Copy the address bar on desktop, or use Share then Copy link in the app. Watch links, youtu.be short links, Shorts and live links all work.
Paste the link into the box on the page. The video's own caption track is read when it has one; when it has none, the audio is run through speech recognition.
The transcript comes back with a timecode on every line. Copy it as plain text, download it as .txt, .srt, .vtt or .doc, or mark a line and export it as a vertical 9:16 clip.
Two sources: the caption track, or listening
Most YouTube videos already carry a transcript. It is the caption track — subtitles the uploader wrote, or the automatic captions YouTube makes itself — and this YouTube transcript extractor reads it directly. The uploader’s own subtitles are taken first, because a person checked them; failing those, YouTube’s automatic captions in the language the video is spoken in. Reading a track costs no listening, so it works at any length: an hour-long lecture with captions is read just as a one-minute Short is.
Automatic captions have a known weakness: they often arrive as a long lowercase run with little punctuation. The words and their timings are right; the sentences are not marked. That is YouTube’s text, handed over as it is.
Some videos have no caption track at all — Shorts most often. For those this listens instead: the audio goes through speech recognition and comes back punctuated, with every word timed. Listening costs us compute, so it runs up to twenty minutes of video free, forty-five on Studio and sixty on Pro. Past that, up to sixty minutes, it is 1 credit per twenty minutes over, priced up front; anything longer is refused up front rather than part-way through.
Either way you get the video’s own words — the uploader’s subtitles as written, or the speech in the language it was spoken — never one of YouTube’s machine-translated tracks. You can then have it translated by AI into one of 20 languages, or summarised into its main points — included with a monthly or yearly plan, or 1 credit each.
What you can take away, and what YouTube itself offers
YouTube has a transcript panel of its own on many captioned videos — open the description and look for Show transcript. It is made for reading beside the video; there is no button there to download it as a file.
Here the transcript is yours to take. Copy transcript puts the plain text on your clipboard, one line per phrase and no timecodes in the way — YouTube to text in one press. Or download it: .txt for the same plain text, .srt or .vtt for a subtitle file with the timecodes in it, or a .doc that opens in Word with a timecode beside every line.
A misheard word can be fixed before you download. Correct a line on the transcript page and every download after that carries your correction.
One video per link
Any link to one video. The watch link (youtube.com/watch?v=…), the youtu.be short link from the share sheet, a Short (/shorts/…), a live stream’s /live/… link and an embed link all name one video, on youtube.com or m.youtube.com, and extra parameters such as a start time are ignored.
Not a channel, a playlist or YouTube Music. None of those is one video, so there is no single transcript to give; paste the link to the video itself.
Not a private or members-only video. We have no login and use nobody’s session, so a video you can only watch while signed in is one this cannot read. When a link cannot be read, the reason comes back in a sentence rather than as a wait that goes nowhere.
Skim a forty-minute talk in two, then jump to the part that answers your question — the timecode against a line is where it is in the video.
Search the transcript for the phrase, copy the line, and cite the moment rather than paraphrase it.
Download the .srt or .vtt of a video you made and take it into your editor, with the timings already in place.
Mark the lines and export them as a vertical 9:16 clip with captions burned in — the reason this site exists. Start from the home page to see how.
Copy the video’s link — the address bar on desktop, or Share then Copy link in the app — paste it in the box above and press Get transcript. The transcript opens on a page of its own with a timecode on every line. There is nothing to install and no account to make.
Yes, in four formats, each one press: .txt for plain text, .srt and .vtt for subtitles with the timecodes in them, and a .doc that opens in Word. Or copy the whole transcript to the clipboard. None of it asks for an account.
Five transcripts a day per browser, no account and no card — the same five across every transcript tool on this site. A video someone else already read does not count, and neither does a read that fails. Past five, each transcript is 1 credit, and so is each AI summary or translation — or a monthly or yearly plan makes transcripts unlimited (fair use: 50 a day) and includes the summary and translation. The rate card is one page.
Then it is listened to: the audio goes through speech recognition and the transcript comes back punctuated, with word-level timings. That runs up to twenty minutes of video free, forty-five on Studio and sixty on Pro, and up to sixty on any plan at 1 credit per twenty minutes over. A video with a caption track has no such ceiling, because nothing has to be listened to.
Every line carries one. On automatic captions and on a video we listened to, the timings are word by word; an uploader’s own subtitles can be timed a phrase at a time, and then the line keeps the uploader’s timing.
Because it was not in the caption track. YouTube’s automatic captions often come as a lowercase run with few full stops, and we hand over the track as it is rather than inventing sentences. Uploaded subtitles keep whatever punctuation the uploader wrote, and a video we listen to comes back punctuated.
Usually the language the video is spoken in. The extractor reads the uploader’s own subtitles in the language they wrote them, YouTube’s automatic captions in the spoken language, or listens to the audio as it is — it never takes one of YouTube’s auto-translated tracks. If you want it in another language, an AI translation into one of 20 languages — included with a monthly or yearly plan, or 1 credit — downloads as .txt or .srt.
Yes — a /shorts/… link and a /live/… link are one video like any other. Shorts often have no caption track, so they are usually listened to.
No. One link, one video, one transcript — paste each video’s own link.
The transcript is cached against the public video, not against you, so the next person who pastes the same link gets it free. What we hold about you is what counting the free five needs. Full list in the privacy policy.
All of them. TikTok, Instagram reels and Facebook video have pages of their own, and a video or audio file on your computer goes to video to transcript. Every tool is listed on one page.