How to get the transcript of a YouTube video (and make notes from it)

Every YouTube video with captions already has a transcript; YouTube just makes it awkward to find and almost impossible to take anywhere. You can read it in YouTube's own "Show transcript" panel, or paste the link into a tool such as vidstill's transcript tool and get the same text as something you can search, copy and export, with timestamps. This guide covers both, what to do when a video has no captions at all, and how a transcript plus a summary can replace watching a long video twice.

YouTube's built-in transcript

YouTube's transcript panel exists, and for a quick look it is fine. Finding it is the first problem.

On desktop:

Open the video and click ...more at the end of the description to expand it.
Scroll to the bottom of the expanded description and click Show transcript. On some layouts the same option sits in the ⋯ menu next to the Share and Save buttons.
The transcript opens in a panel to the right of the video, one caption line per row with a timestamp. Clicking a row jumps the video to that moment.

On mobile:

Tap the video's title to open the description sheet.
Scroll down to the Transcript section and tap Show transcript.
The lines appear inside the sheet, with timestamps.

The panel only appears when the video has captions, whether the creator uploaded them or YouTube generated them automatically. If there is no caption track, there is no panel.

What YouTube's panel cannot do

No export. There is no button to save the text as a file, and no copy button. On desktop you can drag-select the panel and copy, but you get one caption fragment per line, timestamps mixed in, and the panel loads in chunks, so a long video needs several passes.
No search box. The panel has no search of its own. On desktop your browser's Find sometimes catches the loaded part; on mobile it does not work at all.
Copying on mobile is painful. Text selection in the description sheet grabs a line or two at a time, the sheet scrolls under your finger, and the timestamps come along whether you want them or not.
Caption fragments, not paragraphs. Captions are timed to a few seconds each, so the panel reads as a stack of half-sentences rather than prose.
Nothing for videos without captions. The panel can only show a track that already exists.

For reading along while a video plays, the panel does the job. For anything you want to keep, quote, search or study from, you need the text out of YouTube.

Getting a transcript you can use

Copy the video's link. From the address bar on desktop, or Share → Copy link in the app. youtu.be and Shorts links work.
Paste it into the transcript tool and press Go. The card shows the title, channel and length.
Open the Transcript tab. If the video has captions, the text appears within a second or two, joined into readable paragraphs with the time each one starts. The panel tells you whether it came from the creator's captions or YouTube's auto-generated ones; the creator's are preferred when both exist.
Use it. Toggle timestamps on or off, press Copy, or export a file. Click a timestamp to see where in the video that line is spoken.

Copying to the clipboard, with or without timestamps, is free, as is a plain TXT export. SRT and VTT subtitle files and a Markdown export with timestamps are Pro features; they matter if you want captions for a video editor or a file that drops straight into Notion or Obsidian.

When a video has no captions

Some videos have no caption track: the creator disabled captions, the video is too new for YouTube to have processed it, or the audio defeated YouTube's recogniser. The built-in panel simply does not appear. In vidstill, the transcript tab falls back to speech-to-text: it fetches the audio and transcribes it, with timestamps, usually in under a minute. Two limits apply. Speech-to-text only works on videos under 30 minutes, on every plan, because transcribing audio costs far more than reading captions, and it needs actual speech; music videos and ambient footage come back empty or garbled. For a long video with no captions, the honest answer is that no tool gets you a transcript quickly; ask the creator to turn captions on.

Turning a transcript into notes

A transcript of a one-hour lecture is nine or ten thousand words. Reading it is faster than watching, but it is still not notes. The same card has a Summary tab that writes from the transcript in three styles, each written separately rather than the same text at three lengths:

TL;DR. A paragraph on what the video is about, what it concludes and who it is for. Read this first to decide whether the video deserves more of your time.
Key points. The main claims, steps or arguments in the order they come up. This is the style that most resembles the notes you would have taken by hand.
Chapters. The video divided into sections, each with a heading, an approximate start time and a few sentences. For a long interview or podcast, this is the map you use to jump to the part you need.

The workflow that replaces rewatching is: summary first, transcript second, video last. Read the key points to get the structure, search the transcript for the exact wording of anything you want to quote or check, and click the timestamp to watch only that thirty seconds. The summarizer works on any public video; long videos with captions are fine, while uncaptioned ones are capped at 30 minutes by the speech-to-text step.

What people use transcripts for

Lectures and courses. Turn a recorded lecture into text you can annotate, paste into a flashcard tool, or search the night before an exam.
Interviews and panels. Pull exact quotes with times. A quote with a timestamp can be checked by anyone, which is the difference between a claim and a source.
Podcasts published on YouTube. Many exist mainly as YouTube uploads with auto-generated captions; the transcript lets you skim a two-hour episode in ten minutes.
Quoting accurately. Writers citing a video get the speaker's actual words, not what they remember hearing.
Accessibility. Read a video where you cannot play sound, when the speaker is hard to follow, or when reading is simply easier for you than listening.

How accurate is it

Creator-written captions are the most reliable source and are used whenever they exist. YouTube's auto-generated captions and speech-to-text are both good with clear speech from a single speaker and both stumble in the same places: proper names, technical terms, acronyms, heavy accents, crosstalk, and speech over music. Neither adds reliable punctuation. Treat any line you plan to quote as a lead, not a fact, and click its timestamp to hear it before you rely on it. Summaries inherit whatever errors the transcript has, so a misheard name in the captions becomes a misheard name in the key points.

What does not work

Private and unlisted-with-sign-in videos. The tool sees only what a signed-out visitor sees.
Age-restricted videos and live streams. A stream's replay works once YouTube publishes it.
Playlists and channels. Paste one video link at a time.
Translation. You get the language spoken, or the caption languages the creator uploaded. Nothing is translated.
Videos over 30 minutes without captions. Speech-to-text stops there on every plan. Guests and free accounts are also limited to 30-minute videos for captions; Pro raises that to four hours.

A note on other people's words

Reading a transcript, studying from it and quoting short passages with attribution is ordinary use. Publishing the full text of someone else's video, or presenting their words as your own, is not. Transcripts and summaries live on a public page for each video; sign in to keep yours private. Creators can ask for a page to be removed through the DMCA page. If you also want the audio to listen to offline, the MP3 option is on the same card, and the download guide covers the video itself.

FAQ

Is there a free way to get a YouTube transcript? Yes. YouTube's own panel is free but cannot export. vidstill gives guests a few transcripts and summaries a day with copy and TXT export free; a free account raises the allowance.

Can I get the transcript with timestamps? Yes. Every paragraph carries its start time, and a toggle removes the timestamps when you want clean prose. The Markdown and subtitle exports keep them.

Why does my transcript have no punctuation or wrong names? It was auto-generated, either by YouTube or by speech-to-text. Both infer sentence breaks and guess at names. If the creator later uploads their own captions, the transcript improves.

Does this work on iPhone and Android? Yes, in any browser. Copy the text into Notes, or export a TXT file to the Files app on iPhone or the Downloads folder on Android. It is far easier than selecting text in the YouTube app.

Can I get a transcript of a TikTok, Instagram or X video? Yes. Those platforms have no caption tracks to read, so the audio is transcribed with speech-to-text, which takes up to a minute and works on videos under 30 minutes.