How to Get a YouTube Transcript (Any Video, Free)

Nearly every YouTube video already has a transcript — YouTube generates one automatically and shows it on the watch page, four clicks away, no extension required. The catch is what most guides skip: the copy button doesn't exist for videos you don't own, and the text you get has a predictable error pattern you need to fix before using it anywhere.

Any video: the Show transcript button

This works for any public video — yours or anyone else's:

1. Open the video on desktop.
2. Expand the description (click ...more under the title).
3. Scroll to the bottom of the description and click Show transcript. On some layouts it also appears in the three-dot menu below the player.
4. The transcript opens in a panel beside the video. Click the three dots at the top of the panel to toggle timestamps off if you just want prose, then select and copy.

Each line in the panel is clickable and jumps the player to that moment — which makes the transcript panel the fastest way to skim a long video for the part you need, even when you don't want the text itself.

On your phone

The mobile app has the same feature, one tap deeper: tap the video title to expand the description, then tap Show transcript at the bottom. The panel works the same way. Copying long passages on mobile is clumsy, though — if the goal is to reuse the text somewhere, switch to desktop.

Your own video: download the .srt

For your own uploads, don't copy from the watch page — YouTube Studio will give you a proper subtitle file, and that's strictly more useful:

1. Open YouTube Studio → Subtitles in the left sidebar.
2. Pick the video, then the caption track (auto-generated or your own).
3. Open the three-dot menu next to the track and download as .srt.

The difference matters: copied panel text is words alone, while an .srt is words plus the exact timing of every line. Timings are what turn a transcript from a reading document into an editing asset — more on that below.

When there's no transcript button

No button means auto-captions were never generated. The usual reasons:

• The upload is fresh. Captions can lag the video by minutes to hours.
• The audio is mostly music, or speech buried under a loud bed.
• The spoken language isn't supported by YouTube's recognition.
• The uploader turned captions off.

In those cases the transcript has to be made, not found: run the audio through a speech-to-text tool, or type it by hand if accuracy is critical. Our guide to writing a transcript of a video compares the three routes and how long each really takes.

Timestamps: keep them or strip them?

Decide based on what the transcript is for — recovering timings later means redoing the work:

• Strip them for anything meant to be read: blog posts, show notes, quotes, summaries, feeding an AI assistant.
• Keep them for anything time-based: captions, finding and cutting clips, or generating visuals that have to land on specific lines. For these, prefer the .srt download over panel text whenever the video is yours.

The errors to fix first

Auto-transcripts are speech recognition, and recognition fails in a predictable pattern: proper nouns, numbers, and jargon — which are also the words that carry the most meaning. A transcript that says "40%" where the speaker said "14%" is worse than no transcript at all.

Cleaning up is fastest in this order:

1. Names, numbers, technical terms. Highest error rate, highest cost when wrong.
2. Punctuation and paragraphs. Auto-captions arrive as one long run of text.
3. Filler words — remove for reading, keep for captions so the text matches what viewers hear.

What a transcript is actually good for

Most people grab a transcript to quote it or summarize it. Fine — but the timed version is capable of much more:

• Skimming. Read a 40-minute video in four minutes; click a line to jump there.
• Repurposing. Turn one video into posts, threads and articles by working from text instead of scrubbing a timeline.
• Clip selection. Timestamps take you straight to the moment worth cutting.
• Generating visuals. The one most people miss: an .srt is a complete list of what is said and exactly when — that is a shot list.

That last one is where a transcript stops being paperwork. Ahacut reads an .srt and generates a full-length motion-graphics b-roll track — animated text, numbers, charts — synced to every line, ready to drop onto your timeline in CapCut, Premiere or Final Cut. If you just downloaded your transcript from YouTube Studio, you're already holding the input.

Frequently asked questions

Related: How to write a transcript of a video · How to add b-roll to videos automatically · Turn a transcript into motion graphics