The Fastest Way to Get a YouTube Transcript
To get a YouTube transcript, open the video, expand the description by clicking or tapping "...more," scroll to the bottom of it, and hit "Show transcript." A panel opens with every spoken line and its timestamp. This works on desktop browsers and in the iOS and Android apps, it's free, and it needs nothing installed, as long as the video has a caption track.
The difference between devices is what happens next. On desktop you can select the text and copy it. In the mobile app the panel is read-only, so you can follow along but you can't select or copy the text. On a phone you either open the video in a mobile browser and request the desktop site, or use a tool that fetches the transcript for you. Glasp's YouTube Summary extension does that on desktop, and the Glasp mobile app does the same job on a phone, transcript plus a structured summary with timestamps in a few seconds.
Pick your row:
| If you want to... | Use this method | Time |
|---|---|---|
| Read the transcript once, on desktop | YouTube's built-in "Show transcript" | ~15 seconds |
| Read along on your phone | The app's transcript panel (read-only) | ~15 seconds |
| Copy or save text from a phone | Desktop-site request, or a transcript app | ~1 minute |
| Grab transcripts often, on any device | A browser extension or the Glasp app | ~10 seconds |
| Get a transcript without installing anything | Paste the URL into a transcript website | ~30 seconds |
| Pull transcripts in bulk or in code | yt-dlp or a transcript API | Setup, then seconds each |
| Turn the video into a summary and notes | Glasp's transcript tool plus highlighting | ~1 minute |
What a YouTube Transcript Actually Is
A transcript is the full text of everything said in a video, usually broken into short lines each stamped with a timestamp like 4:32. It comes from the video's caption track, and there are two kinds of caption track.
The first is creator-uploaded captions. Some channels write or professionally transcribe their own captions, and these are the most accurate because a human checked them. The second, far more common, is auto-generated captions. YouTube runs the audio through automatic speech recognition and produces captions on its own, which is why almost every video has something you can pull, even when the creator never touched captions.
This matters for one reason: YouTube can only show a transcript when a caption track exists. If a video has neither uploaded nor auto-generated captions, which happens with some music, very short clips, or unusual languages, there's nothing to display. Creators can also delete or unpublish their video's caption tracks, which takes the transcript with them. In either case you need a speech-to-text tool that transcribes the audio itself rather than reading an existing caption track.
Why bother with text at all when you could just watch? Speed and searchability. Marc Brysbaert's 2019 meta-analysis in the Journal of Memory and Language, which pooled 190 studies covering 18,573 participants, put average adult silent reading at 238 words per minute for non-fiction and 260 for fiction. Presented speech, meaning the lectures, monologues and talking-head explainers most of YouTube is made of, runs closer to 150 to 185 words per minute in the classic Tauroza and Allison measurements. Spontaneous conversation is faster than that, so the gap narrows on a podcast, but for the instructional videos people actually want transcripts of, reading is about 1.4 times quicker before you skip a single word of filler. Text is also searchable, quotable, and easy to save, which video simply isn't. For a deeper look at that trade-off, see our guide on how to learn from YouTube.
Method 1: YouTube's Built-In Transcript on Desktop
YouTube's own transcript feature is free, accurate to whatever the caption track contains, and always available on desktop. It's the method to reach for first.
On youtube.com in any browser:
- Open the video and let it start playing.
- Scroll down below the video to the description box.
- Click "...more" to expand the description fully.
- Scroll to the bottom of the expanded description and click "Show transcript."
- A transcript panel appears on the right, with every line and its timestamp.
- To copy clean text, click the three-dot menu at the top of the panel and select "Toggle timestamps" to hide the times, then drag to select all the lines and copy.
Step 6 is the one most people miss, and it's the difference between a usable paragraph and a wall of text interrupted by a number every eight words. Two other things live in that same menu. It's where you switch caption language when a video has more than one track, which is the fastest way to pull an original-language transcript from a video you're watching with translated captions. And clicking any line in the panel jumps the player to that moment, which makes the transcript a usable table of contents for a long video even if you never copy a word of it.
The built-in transcript is perfect for a one-off read. Its limits show up when you do this often: there's no summary, no way to keep the text once you close the tab, and copying strips it out of any system where you'd actually use it later.
Method 2: How to Get a YouTube Transcript on iPhone and Android
The transcript panel does exist in the YouTube mobile apps, on both iOS and Android, and it has since 2022. Guides that tell you the button is missing on mobile are out of date. Google's own support page for viewing transcripts gives identical steps for desktop, Android and iOS.
In the YouTube app:
- Open the video and tap the title, or the small arrow next to it, to expand the description.
- Scroll down through the description until you reach the transcript section.
- Tap "Show transcript."
- The transcript opens as a scrollable list of timestamped lines that auto-scrolls and highlights the current line as the video plays. Tap any line to jump there.
Reading it is easy. Getting the text out of it is not. In the apps the transcript is read-only: you can tap a line to jump to that moment, but you can't select or copy the text, and the panel's own menu offers only a caption-language switch. That gap is what sends most "how to get a YouTube transcript on iPhone" searches looking for a workaround in the first place.
Three ways around it:
- Request the desktop site. Open the video in Chrome or Safari on your phone instead of the app, open the browser menu, and choose "Request Desktop Site." The desktop layout loads, the transcript panel behaves like it does on a computer, and text selection works again. It's fiddly on a small screen but it costs nothing.
- Use a transcript website. Copy the video URL from the app's Share sheet, paste it into a transcript site in your mobile browser, and copy the text from there. Nothing to install, which matters on a locked-down work phone.
- Use an app that fetches the transcript for you. This is the only route that gets you a transcript and keeps it. Note that browser extensions don't help here, because neither iOS nor Android Chrome runs them, so this has to be a real app. The Glasp mobile app pulls the transcript and an AI summary from a shared YouTube link, and anything you highlight saves to your library instead of a clipboard you'll overwrite in ten minutes.
One more phone-specific note: the transcript button lives inside the description, and on mobile the description panel is easy to mistake for the comments section. If you're scrolling and only seeing comments, back out and tap the video title specifically.
Method 3: YouTube Transcript Chrome Extensions
A browser extension puts a transcript, and usually a summary, one click away on every video. That's the right tool if you watch YouTube to learn rather than just once in a while, and it's a desktop-only answer.
Once installed, an extension adds a button near the video. Click it and a panel opens alongside the player with the full transcript, clickable timestamps that jump the video to that moment, and often an AI-generated summary of the key points. The good ones let you pick which model writes the summary and let you save the result instead of losing it when you close the tab.
This is the category Glasp's YouTube Summary sits in, and it adds the piece most transcript tools skip: what happens after. You get the transcript and summary, but you can also highlight the exact lines that matter, and those highlights save to your library next to everything else you've read. The video stops being a tab you close and becomes a note you can search months later. We cover the full workflow in how to summarize YouTube videos, and compare the field in the best YouTube summarizer tools.
A few things to check before you trust any extension:
- Does it keep your transcripts, or just display them once? Display-only tools leave you copy-pasting.
- Does it work on your browser? Most support Chrome and Edge; fewer support Safari or Firefox.
- Does it summarize, or only transcribe? For learning, a summary plus the raw transcript beats either one alone.
- What does it cost at volume? Many are free for a handful of videos a day and paywall heavy use.
- What does it do with the video URLs you feed it? Anything unlisted or private deserves a closer read of the privacy policy.
Method 4: Free YouTube Transcript Generator Websites
If you can't or don't want to install anything, a transcript website is the simplest path. You paste the video URL into a box on the site, and it returns the transcript as plain text you can copy or download.
The flow is the same everywhere:
- Copy the YouTube video URL from the address bar or the Share button.
- Open a transcript site and paste the URL into its field.
- Click the button to fetch the transcript.
- Copy the text, or download it as a
.txtor.srtfile.
These sites are handy on a locked-down work computer, on a phone, or on a borrowed device where you can't add extensions. The trade-offs are real, though. They pull the same auto-generated captions YouTube already has, so they're no more accurate than the built-in transcript. Many wrap the text in ads or push you toward a paid plan for longer videos. Some quietly cap how long a video they'll process. And because you're handing a third-party site the URL, they aren't the place for anything private or unlisted. Treat them as a quick grab-and-go, not a workflow you build on.
Method 5: Developer Tools and the Official API
For bulk work, research, or anything you want to automate, command-line tools and libraries pull transcripts in code. This section is for developers and power users; skip it if a click is all you need.
Start with what the official API can't do. YouTube's Data API v3 has a captions.download endpoint, and people reasonably assume it's the supported way to fetch any video's transcript. It isn't. Google's documentation states the method "requires the user to have permission to edit the video," so it only works on videos you own or manage. It also costs 200 quota units per call against a default allowance of 10,000 units per day. That single restriction explains why every general-purpose transcript tool, including the ones in this article, reads the caption track some other way. If you own the videos, the official API is the right answer and the only one with a stability guarantee. If you don't, it's a dead end.
yt-dlp is a widely used open-source downloader that can fetch caption files without downloading the video. To grab the English auto-generated subtitles as a readable subtitle file:
yt-dlp --write-auto-subs --sub-langs "en" --convert-subs srt --skip-download "VIDEO_URL"
Use --write-subs instead of --write-auto-subs when you want only the creator's manually uploaded captions, which are cleaner when they exist. Pass both flags together and yt-dlp takes the human captions where they exist and falls back to auto-generated ones where they don't.
youtube-transcript-api is a Python library that returns a video's transcript as structured data with no API key and no browser. It's excellent for local prototypes and one-off analysis. There's one large caveat, and the library's own README states it plainly: YouTube blocks most IP ranges known to belong to cloud providers, so code that runs fine on your laptop throws RequestBlocked or IpBlocked the moment you deploy it to AWS, GCP, Azure, or a VPS. Teams running this at scale end up on rotating residential proxies or a managed transcript API, both of which cost money. For personal use on your own machine, it works well.
| Tool | Cost | Best for | Watch out for |
|---|---|---|---|
Official captions.download | Free within quota | Videos you own or manage | Needs edit permission; 200 units per call |
| yt-dlp | Free | Saving subtitle files, offline archives | Command line only |
| youtube-transcript-api | Free | Local scripts and research | Cloud IPs get blocked |
| Managed transcript API | Paid | Production apps, bulk jobs | Ongoing cost |
Whichever route you take, you'll end up with one of four file formats, and they aren't interchangeable:
| Format | What it holds | Use it when |
|---|---|---|
.txt | Plain text, no timing | You want to read, quote, or paste into notes |
.srt | Numbered lines with start and end times | You need timestamps or are re-uploading captions |
.vtt | Web-native captions, supports styling | You're embedding captions in a web player |
.json | Structured lines with offsets | You're feeding a script or a model |
Once you have the raw text in code, the interesting part is what you do with it: feed it to a model for summarization, run search across a channel's back catalog, or build your own study notes. Our piece on turning YouTube into study notes covers that downstream step.
How Accurate Are YouTube Transcripts?
Auto-generated transcripts are good enough to read and search, but not good enough to quote blindly. Knowing where they break saves you from citing a word the speaker never said.
For clear English audio with a single speaker and no background noise, modern speech recognition lands in the low-to-mid nineties for word accuracy, and slips into the high eighties as soon as the audio gets conversational. Real videos are often messier than either case. A 2024 paper by Fathallah, Bhole and Staab on improving captions for deaf and hard-of-hearing viewers measured a 23.07 percent word error rate on YouTube's own auto-captions, roughly one word in four, across 52 short clips the authors deliberately picked to include hard cases like ambient noise, homophones and fast speech. That's a stress test rather than an average, but it shows how far the floor can drop. Accuracy falls in a predictable order:
| Condition | Rough accuracy | Why it drops |
|---|---|---|
| Studio audio, one clear speaker | 90 to 95% | Ideal case for the model |
| Conversational English, decent mic | 85 to 90% | Overlap, filler, pace |
| Multiple speakers or crosstalk | Noticeably lower | The model can't separate voices |
| Music or noise near speech level | Sharp drop | Music masks speech as signal-to-noise falls |
| Strong accents or non-English audio | Lower and variable | Less training coverage |
That last row isn't a hunch. Rachael Tatman's 2017 paper "Gender and Dialect Bias in YouTube's Automatic Captions," presented at the first ACL Workshop on Ethics in NLP, tested YouTube's captions across two genders and five dialect groups using videos where speakers state their own language background. She found reliably lower accuracy for women and for speakers from Scotland, and in no dialect group were the captions reliably better for women than for men. If you're transcribing a speaker whose accent isn't General American, expect more cleanup.
The errors also cluster around exactly the words you most want to get right: proper names, brand names, and technical terms, because the model substitutes a common word that sounds similar. It's worth knowing how low that bar sits against the accessibility standard. Captioning guidance aligned with WCAG treats 99 percent accuracy as the target for viewers who depend on captions, and puts typical machine captions somewhere in the 80 to 95 percent band, which is to say below the line. So when you pull a quote from a transcript, spot-check the timestamp against the audio before you publish it. Creator-uploaded captions, when a channel provides them, don't have this problem because a human wrote them.
How to Turn a YouTube Transcript into Notes You Remember
Getting the transcript is the easy part. The reason most people forget nearly everything they watch isn't a lack of transcripts, it's that the text goes nowhere. A transcript you copy and close is as forgotten as the video itself.
Turning a transcript into knowledge takes three steps:
- Highlight the lines that matter. A 40-minute video might have five sentences worth keeping. Mark them instead of saving the whole wall of text. Active selection is itself a learning act: deciding what's important forces you to process it rather than let it wash past.
- Summarize in your own words. Rewriting the key points in your own language is one of the eight generative learning strategies Fiorella and Mayer identify, and the rewriting is where the work happens. AI can draft the summary from the transcript, but the retention comes from you editing it.
- Save it somewhere searchable. A highlight you can't find again might as well not exist. When your video highlights live in the same searchable library as your article highlights and book notes, each new video connects to what you already know.
This is the workflow Glasp is built for. You get the transcript and an AI summary from YouTube Summary, highlight the exact lines you want to keep, and everything lands in one library you can search and question later with Glasp's AI chat. It works the same way Glasp's web highlighter does for articles and PDFs, so videos, articles, and books become one connected body of notes instead of scattered tabs. If retention is your real goal, start with how to remember what you read.
Frequently Asked Questions
How do I get a transcript of a YouTube video for free?
Open the video, expand the description by clicking or tapping "...more," scroll to the bottom, and select "Show transcript." A panel appears with the full timestamped text. This is completely free, needs no software, and works on desktop browsers and in the mobile apps. On desktop you can select and copy the text directly; on a phone you'll need to request the desktop site in your browser, or use a transcript website or a summarizer app, because the panel is read-only in the app.
How do I get a YouTube transcript on my iPhone?
Tap the video title to expand the description, scroll down, and tap "Show transcript." Reading works fine, and the steps are the same on Android. Copying doesn't, because the panel is read-only: its only menu option is switching the caption language. To get the text out, open the video in Safari or Chrome, choose "Request Desktop Site" from the browser menu, and use the desktop transcript panel, or paste the video URL into a transcript website or an app that fetches it for you.
Can I copy a YouTube transcript without the timestamps?
Yes, on desktop. Open the transcript panel, click the three-dot menu at the top of it, and choose "Toggle timestamps." The times disappear and you're left with clean prose you can select and copy in one pass. Without that step you get a number interrupting the text every few words, which is the main reason pasted transcripts look unusable. The toggle isn't available in the mobile apps.
Why can't I see the "Show transcript" button?
Usually the video has no caption track, so there's nothing to display. YouTube only shows a transcript when captions exist, whether uploaded by the creator or auto-generated. Creators can also delete or unpublish their caption tracks, which removes the transcript too. And on mobile the button sits inside the description panel, which is easy to miss if you've opened the comments instead.
How do I download a YouTube transcript as a text file?
The easiest no-code route is a transcript website that offers a download button for .txt or .srt files after you paste the URL. Developers can use yt-dlp with --write-auto-subs --convert-subs srt --skip-download to save the subtitle file directly. You can also copy the text from YouTube's built-in transcript panel on desktop and paste it into any document.
Is there an official YouTube API for transcripts?
Yes, but it probably won't do what you want. The Data API v3 captions.download method requires permission to edit the video, so it only returns captions for videos you own or manage, and it costs 200 quota units per call against a 10,000-unit daily default. For other people's videos, tools use yt-dlp, a library like youtube-transcript-api, or a paid managed API instead.
Are YouTube auto-generated transcripts accurate?
Mostly, but not perfectly. Clean single-speaker English audio lands in the low-to-mid nineties for word accuracy, while messier real-world audio drops much further; a 2024 study measured a 23 percent word error rate on a deliberately difficult sample of YouTube auto-captions. Accuracy falls with background music, multiple speakers, and accents, and research by Rachael Tatman found systematically lower accuracy for women and Scottish speakers. Proper names and technical terms are the most common casualties, so verify a quote against the audio before publishing it.
Can I get a transcript for a video in another language?
Yes, if the video has captions in that language. YouTube auto-generates captions for many languages, and the three-dot menu inside the desktop transcript panel lets you switch caption tracks when more than one exists. You can also translate captions, though machine translation adds its own errors on top of the transcription.
Five methods, one decision: YouTube's built-in panel on desktop, the same panel on your phone, a browser extension, a paste-a-URL website, or developer tools like yt-dlp. Pick the one that matches your device and how often you need this.
Then do the part that actually decides whether the video was worth the time: highlight the handful of lines worth keeping, put them in your own words, and store them where you'll find them again. Start with Glasp's YouTube Summary on desktop or the Glasp app on your phone, and turn the videos you watch into knowledge that compounds instead of vanishing when you close the tab.
