YouTube Transcript Extractor

Extract existing captions from a public YouTube video. Copy the transcript or download it as TXT, Markdown, SRT or VTT.

Read sends the video id to this site's server. The server brings back the caption text this public video already has.

The captions already attached to the video

Lessons, interviews, and talks often already have captions. Paste a public link and you get that text, not a new script written by listening again.

  • Manual captions are used when they exist. Automatic captions are used when they are the only ones, and the track list shows which kind. Do not treat an automatic track as a script someone has already fixed. Names and sentence breaks are often still the machine’s.

  • Lines stay in time order. Read a few on the page, check the episode and the language, then download. One video can carry several languages. Do not assume the first track is the one you wanted.

  • The link only needs a video id. The server uses that id to fetch public captions. You do not sign in to YouTube. A private, deleted, or unavailable video stops with an error instead of returning something that looks like a script and is empty.

  • No caption track, or an empty one, and the page says nothing was read. Sound is not captions. If the creator never turned them on, there is no text, and none is invented. Too many reads in a short time pause the page. A miss is not packaged as an empty file.

Plain text for reading, a timeline for the edit

The same captions want different files depending on whether you are quoting them or cutting picture to them.

  • For notes, use TXT or Markdown so the timing does not chop the sentences. For a cut on the timeline, use SRT or VTT so each line still knows where it sat.

  • Mistakes in automatic captions stay in the original text. Nothing is quietly rewritten before download. Edit them in the subtitle tool you already use. This page is how the original text gets out.

  • Comments are a CSV on another page. Lines such as “0:00 Intro” in the description are on the chapter page. Spoken lines, viewer comments, and a chapter list stay three documents even when their times are close.

How to use it

  1. Paste the link of a public video. A private, deleted, or unavailable video stops with an error.
  2. The server uses the video id to fetch captions that are already public. Manual captions are used when they exist. Automatic captions are used when they are the ones available.
  3. Read a few lines on the page, then copy them or download TXT, Markdown, SRT, or VTT.

After downloading SRT or VTT, use zer.ai Subtitle Editor to correct the text or adjust the timing. If your video editor needs another format, use zer.ai Subtitle Converter.

Limits

  • Only captions already on a public video are read. If there is no text, the page says so and does not turn a blank result into a script.
  • The video id is sent to this site so the server can fetch that public caption text. You do not sign in to YouTube.
  • Too many reads in a short time pause the page. Try again later. A failed read is not handed to you as an empty script.

FAQ

Can I download the caption text?

Yes. When a public video already has captions, you can copy them or download plain text, Markdown, or timed SRT and VTT. The timed files are the ones to take into an edit or a proofread.

What if the video has no captions?

The page says nothing was read. A blank result is not turned into an article and is not treated as a successful extract. If automatic captions exist, those are used, and the track list shows which kind you received.

Related tools