Back to blog

How to Download YouTube Captions & Transcripts (2026 Guide)

If a YouTube video already has captions, either uploaded by the creator or auto-generated by YouTube, you don't need to transcribe it from scratch. You just need to get that existing caption data out as a usable text file. That's a different problem from generating a transcript, and most guides on the internet blur the two together.

This guide is specifically about extracting captions that already exist on a video. If the video has no captions at all (common on smaller channels, Shorts, and non-English content), skip to the last section, you'll need actual AI transcription, not extraction.

Why People Need This

A few common reasons someone wants a YouTube video's captions as a downloadable file:

  • Research and note-taking, pulling quotes or key points from a long lecture, interview, or conference talk without re-watching it.
  • Repurposing, turning an existing video (yours or one you have rights to reference) into a blog post, summary, or quote graphic.
  • Accessibility archiving, keeping a text record of a video in case it's taken down or region-locked later.
  • Translation prep, getting the English caption track as a base to translate into another language.

None of these require re-transcribing the audio, the caption track already exists. The trick is just getting it off YouTube's player and into a plain text or SRT file.

Method 1: YouTube's Built-In Transcript Panel (Free, No Tools Needed)

The simplest method works entirely in the browser, no downloads or extensions required:

  1. Open the video on YouTube (desktop web, not the app).
  2. Click the "...more" button below the video title, or the three-dot menu.
  3. Select "Show transcript."
  4. A scrollable transcript panel opens on the right, with timestamps next to each line.
  5. Click the three-dot menu inside the transcript panel and toggle off timestamps if you just want clean paragraph text.
  6. Select all the text (Ctrl+A / Cmd+A inside the panel) and copy-paste it into a document.

Limitations: This gives you plain text, not a proper timed .srt/.vtt file, so it's fine for reading or quoting, but not for re-uploading as a caption track somewhere else. It also only works if the video has a transcript available (auto-generated or uploaded) and you can't select transcripts for videos in some restricted/age-gated states without extra steps.

Method 2: Downloading an Actual .srt/.vtt File with yt-dlp

If you need a real timed caption file, not just pasted text, the free open-source tool yt-dlp is the standard method:

yt-dlp --write-auto-sub --sub-lang en --skip-download "https://www.youtube.com/watch?v=VIDEO_ID"

This downloads only the English auto-generated subtitle track (as .vtt) without downloading the video itself. Swap --write-auto-sub for --write-sub if you specifically want creator-uploaded captions rather than YouTube's auto-generated ones (not every video has both).

Why this is the more "real" method: it gives you an actual timed subtitle file you can open in a text editor, convert to .srt, or re-upload elsewhere, not just a copy-pasted blob of text.

The catch: yt-dlp requires comfort with a command line, needs to be kept updated (YouTube changes its extraction methods often enough that an outdated version starts failing), and, for JavaScript-heavy pages, sometimes needs a JS runtime like Deno installed alongside it to parse newer player responses correctly. If that sentence made your eyes glaze over, Method 1 or a browser extension is the more realistic option.

Method 3: Browser Extensions

Several browser extensions (search "YouTube transcript downloader" for your browser) add a one-click "download transcript" button directly on the YouTube watch page. These are convenient but come with real caveats:

  • Quality and maintenance vary wildly, check recent reviews before installing, since YouTube's page structure changes periodically and abandoned extensions break silently.
  • Any extension that can read page content technically can see everything else on pages you visit too, stick to extensions with a narrow, clearly stated permission scope and real user reviews, not just install count.
  • Most only export plain text, same limitation as Method 1.

Method 4: Third-Party Web Tools

Sites like youtubetotranscript.com-style tools let you paste a YouTube URL and get a transcript back without installing anything. Fine for occasional one-off use. Same tradeoffs as extensions apply, verify the tool is still actively maintained, since these break every time YouTube adjusts its caption delivery, and be cautious about pasting URLs of anything private or unlisted into a third-party site.

Important Limits of All Four Methods

Every method above is extraction, not transcription, they only work if:

  1. The video already has captions. No captions on the source video = nothing to extract, full stop.
  2. You're allowed to use them. Auto-generated YouTube captions are famously imperfect, no reliable punctuation, awkward line breaks, and real accuracy drops with accents, cross-talk, technical vocabulary, or background music. Fine for a rough read; not something you'd want to republish as your own accurate transcript without cleanup.
  3. The video isn't otherwise restricted, age-gated, region-locked, or private videos add friction to every method above, extension and command-line alike.

If any of those don't hold, no captions exist, the auto-captions are too messy to trust, or you need a clean .srt/.vtt you can actually publish, extraction tools can't help. That's a transcription problem, not an extraction problem.

When You Actually Need Transcription, Not Extraction

If the video has no usable captions, or the auto-captions are too rough to trust for something you're publishing, the fix is running the audio through a real transcription model rather than trying to scrape text that doesn't exist yet.

This is the point where a tool like Tapescribe is the more direct path: paste the YouTube (or Vimeo/Loom) URL, and it returns a proper transcript, timed SRT/VTT captions, and auto-detected chapter markers, typically in a few minutes, without a command line or extension. It's flat $1/video with no subscription, and the free tier includes 3 videos with no card required, so it's a reasonable way to test accuracy on your specific content before committing to it for a whole back-catalog.

The practical rule of thumb:

  • Video already has decent captions, you just need the text → use Method 1 (browser transcript panel) for quick reading, or Method 2 (yt-dlp) if you need a real timed file.
  • No captions exist, or the existing ones are too rough to trust → transcribe it properly instead of trying to extract text that isn't there.

Frequently Asked Questions

Can I download captions from a YouTube video I don't own?

Technically yes, all four methods work on any public video with an existing caption track, but what you're allowed to do with that text is a separate question. Extracting a transcript for private research, note-taking, or quoting with attribution is very different from republishing someone else's video content as your own page. If you're building content around video you own, transcripts are genuinely useful for search visibility, which is covered in the video transcription and YouTube SEO guide.

Why does the transcript panel option not appear on some videos?

The "Show transcript" button only shows up when a caption track actually exists on that video. Creators can disable auto-captions entirely, and YouTube sometimes fails to generate them for very short clips, heavily music-based audio, or languages its auto-caption system doesn't support well. If the button is missing, there is nothing to extract and you'll need to transcribe the audio instead.

Can I convert the downloaded .vtt file into an .srt?

Yes. The .vtt file yt-dlp produces is plain text with timestamps, so it can be converted to .srt with a small script, a text editor, or any of the free format converters online. That said, if the underlying captions came from YouTube's auto-generation, converting the format doesn't fix the missing punctuation or awkward line breaks, so you're still starting from rough source text.

Do the timestamps in a downloaded transcript match the video exactly?

Timed files from yt-dlp keep YouTube's original cue timings, so they line up with the video as published. Copy-pasted text from the transcript panel keeps timestamps only if you leave that toggle on, and pasting into a document often mangles the formatting. If your goal is captions you'll actually display to viewers rather than read yourself, timing and line length matter a lot more, which is what caption best practices walks through.

Related Reading