How to Get a YouTube Transcript: 5 Methods to Extract Text

A YouTube video that has captions already carries its text, so extracting it is a copy operation rather than a transcription job. YouTube stores one caption track per language per video, written either by the creator or by YouTube's own speech recognition. Every method below reads that same track. What separates them is how many videos you can run through at once, which file formats come out, and how much setup you are willing to do first.

Five methods, in order of setup cost: YouTube's own transcript panel, YouTube Video Transcript in the browser, a Chrome extension, the yt-dlp command line, and a bulk download across a whole playlist or channel. Each section below gives the steps and the point at which that method stops working.

How do I copy a YouTube transcript without installing anything?

Open the video on desktop YouTube, expand the description, and click Show transcript. A panel opens beside the video with every caption line in it, and you can select the whole thing and copy it. No account, no extension, no download.

1. Open the video on youtube.com in a desktop browser.
2. Click ...more under the video title to expand the description.
3. Scroll to the bottom of the expanded description and click Show transcript. On some layouts the same option sits in the three-dot menu next to Save.
4. Click inside the panel, press Ctrl+A or Cmd+A to select all, then Ctrl+C or Cmd+C to copy.
5. Paste into a text editor, a Google Doc, or a notes app.

The panel's three-dot menu toggles timestamps on and off. With them on, every copied line arrives prefixed with a time such as 0:14, which is what you want when you plan to cite a moment and what you do not want when you plan to read the thing as prose.

This method covers one video at a time and produces plain text on the clipboard. There is no SRT, no JSON, and no file. By the third or fourth video the open, expand, click, select, copy, paste loop costs more attention than the transcript is worth.

How do I export a YouTube transcript as SRT, JSON, or DOCX?

Paste the video URL into a tool that formats the caption track for you. YouTube's panel only ever hands you plain text. YouTube Video Transcript returns the same caption data as TXT, SRT, JSON, CSV, DOCX, or Markdown, and it is the tool we build.

1. Open the YouTube video transcript homepage.
2. Paste the video URL into the input box.
3. Read the transcript that appears on screen.
4. Sign in with Google to download it as a file.
5. Pick TXT, SRT, JSON, CSV, or Markdown and download. DOCX is built server-side, so it comes through the API rather than the browser picker.

You can check the output before you commit to anything. Paste a URL on the homepage and the first transcript renders on screen without an account at all. Saving a file is what requires the Google sign-in, and signing in comes with 30 transcripts a month at no cost and no card.

The formats are not six wrappers around the same file. SRT numbers each caption block and writes a timecode pair, 00:00:14,120 --> 00:00:19,400, which is what a video editor reads. JSON gives every segment a start and end in seconds, which is what a chunker for a RAG index wants. CSV opens in a spreadsheet with a start,end,duration,text header row. Markdown puts a readable timestamp in bold at the head of each line for Obsidian or Notion. DOCX and TXT drop the timing and keep the words.

Do Chrome extensions work for grabbing YouTube transcripts?

Yes, for one video at a time. Extensions such as Tactiq and Glasp add a transcript button to the YouTube page itself, so you read and copy the text without leaving the video or pasting a URL anywhere. For a person who watches a lot and occasionally wants the words, that saved context switch is the whole value.

The limit is the shape of an extension. It acts on the page you are looking at, so there is no channel and no playlist, and export options are narrower than a dedicated tool's. An extension is a reading aid, not a way to collect a corpus.

How do I download YouTube captions with yt-dlp?

yt-dlp downloads caption files from the command line, free, with no cap on how many. It is genuinely the right answer when you are already comfortable in a terminal and the job is a handful of videos or a script you want to own.

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

That writes a .vtt file next to your working directory, because VTT is what YouTube actually serves. YouTube offers its caption tracks as VTT, TTML, and its own srv1, srv2, srv3, and json3 variants, and SRT is not among them, so passing --sub-format srt will not produce one. Convert on the way out instead:

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

--convert-subs needs ffmpeg on your PATH. Point the same command at a playlist URL and yt-dlp walks every entry on it; add --yes-playlist when the URL is a single video that happens to sit inside a playlist.

What yt-dlp hands back is one caption file per video in YouTube's format. Filenames, the fallback when a requested language track is missing, deduplication across a channel, and conversion into whatever your pipeline reads are all yours to write and to keep working. That is a fair trade for a script you run twice a year and a bad one for a job you run weekly.

How do I get transcripts for an entire playlist or channel?

Paste the playlist or channel URL into a tool that enumerates the videos for you and returns one archive. The first four methods all break at this size for the same reason: each is built around a single video ID, so a 400-video channel means running the method 400 times.

YouTube Video Transcript takes a channel or playlist URL, reads every public video on it, pulls each caption track, and packages the results as a single ZIP in the format you chose. One paste, one file back. This is the case the product exists for, and it is the case where copy-paste and extensions have nothing to offer.

The largest single job YouTube Video Transcript has handled pulled 3,713 videos from one channel, and 16,466 transcripts have gone through the service so far, at a median of 1.9s per transcript. Figures as of September 2026.

For the longer walkthrough of bulk jobs, see our complete guide to downloading YouTube transcripts, and the 2026 roundup of transcript downloaders compares the tools that handle channels against each other.

Which method should I use?

Volume decides it, and format needs break the tie. The table below maps a job to the method that costs the least effort for it.

Which export format keeps the timestamps?

SRT, JSON, CSV, and Markdown carry timing. TXT and DOCX carry only the words. Those are the six formats YouTube Video Transcript exports, and the caption data underneath every one of them is identical.

How accurate is a YouTube transcript?

Accuracy depends on which caption track the video has, not on the method you used to fetch it, because all five methods read the same track. Three kinds of track exist, and they are not close in quality.

Creator-uploaded captions were written or corrected by a person, so they match the script including names and jargon. Auto-generated captions come from YouTube's speech recognition and are usable for clean single-speaker English while dropping or mangling proper nouns, technical terms, overlapping speakers, and strong accents. Auto-translated captions are machine translation stacked on top of machine recognition, so both error rates compound.

Check which one you have before you trust the text. Open the caption menu on the video: a track listed as English is the creator's, and a track listed as English (auto-generated) is YouTube's. For anything you plan to publish, quote, or cite, start from videos in the first group.

What do people do with the extracted text?

The transcript is an input, not an output. The four uses that come up most often are summarising, searching, repurposing, and translating.

Paste a transcript into Claude, ChatGPT, or Gemini and ask for a structured summary, the direct quotes, or a timestamped outline. Chunk and embed a channel's worth of transcripts into a vector database and you can search a body of video by meaning rather than by title; the JSON export exists for exactly this, since each segment carries its own start and end. A 30-minute talk transcribed is enough raw material for a blog post or a newsletter issue. And machine translating the extracted text in one LLM call usually reads better than YouTube's auto-translated caption track, because the model sees the whole transcript instead of one caption line at a time.

What does this cost once the free methods run out?

YouTube's transcript panel and yt-dlp are free at any volume, and nothing we sell beats free. Paid plans start where your own time becomes the expensive part. YouTube Video Transcript sells four tiers, subscriptions only, with no credit packs:

The pricing page has the full breakdown. There is no trial on the paid tiers. The free 30 a month keeps running instead.

What are the limits worth knowing before you pay?

Four cases, stated plainly so you do not find them after paying.

We do not export VTT. The six formats are TXT, SRT, JSON, CSV, DOCX, and Markdown, so a pipeline that specifically wants .vtt should use yt-dlp, whose native output already is VTT. Credits do not roll over: an unused allowance resets on your billing anniversary rather than accumulating, and a lapsed subscription drops back to the free 30 a month. We read the captions YouTube already stores and run no speech recognition of our own, so a video with captions disabled has nothing for us to fetch. And if you are fluent in a terminal and your job is twenty videos once a year, yt-dlp does it for nothing.

The case we do win is the one that made you search: a whole channel or playlist, in a format your next tool can read, without writing or maintaining anything.

Start with thirty videos

Paste a video, playlist, or channel URL on the homepage. The transcript for a single video renders on screen before you sign in. Sign in with Google and you get 30 transcripts a month at no cost, downloadable as TXT, SRT, JSON, CSV, or Markdown, with DOCX through the API, and a playlist or channel arriving as one ZIP. If thirty is not enough, Starter is $9 a month for 1,000.

Frequently asked questions

How do I extract text from a YouTube video for free?

Open the video on desktop YouTube, expand the description with the ...more link, click Show transcript, then select all and copy. That costs nothing and needs no account. yt-dlp does the same job from a terminal, free at any volume. YouTube Video Transcript's free tier covers 30 transcripts a month after a Google sign-in, which is the route to take when you need SRT, JSON, CSV, DOCX, or Markdown rather than plain text on the clipboard.

Can I extract text from a YouTube video without captions?

No. Every transcript tool reads the caption track YouTube already stores against the video, so a video with no track has nothing to fetch. YouTube auto-generates captions for most uploads in supported languages, so a video without creator-written subtitles usually still has an auto-generated track. When a creator has disabled captions outright, the only route left is downloading the audio and running speech recognition yourself with a model such as Whisper.

What is the most accurate extraction method?

Accuracy comes from the caption track rather than the extraction method, because every method reads the same track. Creator-uploaded captions match the script. Auto-generated captions drop or mangle proper nouns, technical terms, and overlapping speakers, and auto-translated captions add machine translation errors on top of that.

Can I extract text with timestamps?

Yes, in four of the six formats YouTube Video Transcript exports. SRT writes a numbered block with a start and end timecode, JSON gives each segment a start and end in seconds, CSV writes start, end, and duration columns, and Markdown prefixes each line with a readable timestamp such as [12:04]. TXT and DOCX carry the words only.

Is it legal to extract text from a YouTube video?

Extracting publicly available captions for personal use, research, accessibility, or note-taking generally falls under fair use in most jurisdictions. Redistributing verbatim transcripts at scale, or using them as training data for a commercial AI product, raises separate copyright questions. Check YouTube's Terms of Service for your specific use case.

How do I extract text from an entire playlist?

Paste the playlist URL into a bulk transcript tool. YouTube Video Transcript reads every public video on the playlist, pulls each caption track, and returns them as one ZIP in the format you picked. yt-dlp does the same from a terminal: point --write-subs --write-auto-subs --skip-download at the playlist URL and it walks every entry, one caption file per video.

Which format should I pick?

TXT for reading or pasting into an LLM prompt. SRT for video editing and subtitle tracks. JSON for developer pipelines and RAG chunking, because each segment carries a start and end in seconds. CSV for spreadsheet analysis. DOCX for sharing in Word or Google Docs. Markdown for Obsidian or Notion, with a readable timestamp on every line.