What this tool does
Paste any public YouTube URL and you get three things on this page: the full transcript with timestamps, a short summary, and a chapter list you can click through.
Most videos already carry a caption track — either one the creator uploaded or one YouTube generated automatically. When that track exists, this tool reads it directly, which takes a second or two. The raw captions are free without an account; clean punctuation, the summary, chapters and downloads need a free account. When a video has no captions, or YouTube refuses to hand them to this site, there is nothing here to read: if you have the audio or video file, the audio transcription tool transcribes the speech itself, priced per minute.
Nothing is emailed to you and nothing is stored on a public page. The transcript appears in the panel above, you copy it or download it, and that is the end of the transaction.
How to get a transcript from a YouTube video
- Copy the video link from the address bar, the Share button, or the mobile app. Full links, youtu.be short links, Shorts links and links with a timestamp all work.
- Paste it into the field above and press Get transcript.
- Read the result. Use the Copy button for the whole transcript, or select a single paragraph and copy just that.
- If you need subtitles or a Word file, create a free account and pick a download: plain text, Markdown, SRT, VTT and Word, all free. For subtitles in another language, take the SRT to the subtitle translator.
One transcript a day of a video up to 15 minutes is free to read and copy without an account, as YouTube's own captions. A free account gives 2 videos a day up to 20 minutes and adds clean punctuation, the summary, chapters, downloads and longer videos.
Summary and chapters, not just a wall of text
A raw transcript of a 40-minute video is roughly 6,000 words, which is not much use if you are trying to decide whether the video is worth watching.
The summary is a two-sentence description of what the video actually covers plus five to eight points that carry the substance. The chapter list is derived from where the topic genuinely changes, so the timestamps point at the start of a section rather than at an arbitrary five-minute interval. Click any chapter to jump the transcript to that point.
If the video already has creator-written chapters, those are kept as-is — the tool does not overwrite information the author gave you.
Who uses this
Students who need the wording of a lecture in a document they can annotate, then turn it into quiz questions or flashcards.
Researchers and journalists quoting an interview, where the exact sentence and its timestamp both matter.
Video editors and podcasters who want an SRT for burn-in captions without paying for a subtitle subscription, and who then correct two or three lines by hand.
People who read faster than they watch and want the content of a 45-minute video in ninety seconds.
Non-native speakers who follow spoken English better on the page than in the ear, especially at speed.
Limits worth knowing before you start
- Private, unlisted and members-only videos cannot be read. If the video needs a login, this tool cannot see it either.
- Live streams work only after they end and the recording is processed.
- Auto-generated caption tracks contain mistakes, especially on names, jargon and acronyms. They are usually 90–95% accurate on clear speech and noticeably worse on heavy accents, poor audio or music.
- Auto-generated tracks carry no punctuation from YouTube. Punctuation and paragraph breaks in the output above are restored by this tool, which is an improvement but not perfect.
- Speaker labels are not available from YouTube caption tracks. For speaker-separated output, use the audio transcription tool on the audio or video file, if you have it; it can label speakers.
- Very long videos (over three hours) are slow. Free transcripts cover videos up to 15 minutes without an account, and the first 20 minutes with one; paid length uses 1 credit per 3 minutes, rounded up per video, with a 2-credit minimum.
Accuracy, and what to do about it
Treat any transcript as a draft you are responsible for. Two habits cover most of the risk.
First, check every proper noun. Names of people, companies, drugs and places are where automatic transcription fails hardest, and a wrong name in a quote is the kind of error a reader notices.
Second, verify a quote against the audio before you publish it. Every paragraph above carries the timestamp it came from, so this takes seconds: click the timestamp, listen to the sentence, move on.
If the audio quality is genuinely poor and you have the audio or video file, the audio transcription tool's speech model will usually beat YouTube's automatic captions by a visible margin. That costs credits because it costs real compute.
Example result
https://www.youtube.com/watch?v=dQw4w9WgXcQ (3:33, auto-captions)
Summary: two sentences on what the video covers, then 6 bullet points. Chapters: 00:00 Opening, 00:43 First verse, 01:16 Chorus… Transcript: 14 timestamped paragraphs, punctuation restored, ready to copy.
Total time: about four seconds, because the video already had a caption track.
How this compares
Where another tool is the better choice, it says so.
YouTube's own transcript panel
Better thereIt is built into the player, it is instant, and it costs nothing at all.
Better hereIt gives you unpunctuated lines in a scroll box with no summary, no chapters, no paragraphing and no file to download. This gives you a readable document, a summary, and SRT, VTT or DOCX.
NoteGPT / YouTube summary extensions
Better thereA browser extension sits next to the video, so summarising what you are already watching is one click.
Better hereNo extension to install and nothing reading your browsing. You also get the full transcript rather than a summary alone, and a real subtitle file rather than text in a sidebar.
Descript
Better thereA genuine editing suite: correct the transcript and the video edits with it, plus studio-grade tooling.
Better hereDescript is a subscription and a download. This is a page you paste a link into, with credits you buy once and no expiry, for when you need the words rather than an editing workflow.
Your questions, answered.
Is this free?
Without an account: 1 video a day, for videos up to 15 minutes, as YouTube's own captions you can read and copy. With a free account: 2 videos a day up to 20 minutes each (up to 30 a calendar month), with clean punctuation, plus the summary, chapters and every download (text, Markdown, SRT, VTT and Word). Paid length uses 1 credit per 3 minutes, rounded up per video, with a 2-credit minimum. The $3 pack covers up to 150 YouTube minutes. Failed runs are never charged.
Do I need an account?
No. Paste a link and read the transcript, 1 video a day. A free account gives 2 videos a day up to 20 minutes and adds clean punctuation, the summary, chapters, downloads and longer videos. There is no newsletter.
Does it work on videos with no captions?
Not on this page: it reads the captions a video already has. If you have the audio or video file, upload it to the audio transcription tool, which transcribes the speech and can label speakers, priced per minute. The same applies when YouTube refuses to hand a video's captions to this site, which it does for some videos.
What about private or age-restricted videos?
Private, unlisted, members-only and deleted videos cannot be read. Age-restricted videos usually fail too, because the caption track sits behind a signed-in request.
Can I get subtitles I can load into Premiere or CapCut?
Yes. SRT and VTT downloads carry millisecond timing and are accepted by Premiere, Final Cut, DaVinci Resolve, CapCut, Descript and every major web player.
Which languages are supported?
Any language the video has a caption track in, which in practice is most of the languages YouTube supports. For audio transcription the underlying speech models cover about 50 languages well, with the best results in English, Spanish, Portuguese, German, French, Hindi and Mandarin.
How accurate is the transcript?
On clear speech, 90–95% word accuracy from YouTube's own captions and slightly better from a current speech model. Accuracy drops on heavy accents, crosstalk, background music and technical vocabulary. Always check names and numbers before quoting.
Are my transcripts published anywhere?
No. Outputs are private to your job and are never rendered on a public URL. The only pages on this site are ones we wrote.
Can I transcribe a playlist or a channel in one go?
Not yet. Bulk runs over a list of links are planned as a credit-priced feature; today you paste one link at a time.
Why is the transcript missing the last few minutes?
That happens when a free account run is cut at the free length of 20 minutes. The panel says where it stopped, and credits finish the full video. Without an account, videos over 15 minutes need a free account or credits.