Which videos will not work?

Any video with captions disabled by the uploader, and any video that is private, unlisted behind a link you did not sign in for, age-restricted, or removed. Music videos and silent footage often have a caption track so sparse that the result is not worth reading. Everything else works, because YouTube auto-generates captions on the vast majority of uploads.

Are the timestamps YouTube's real cue times?

No, and no tool that flattens the caption track can honestly claim otherwise. The timings here are reconstructed from the video's runtime and the length of each line, which is close enough to jump to a quote but not close enough to burn in as subtitles without checking them in an editor first. If you need frame-accurate subtitles, pull the caption file from YouTube Studio.

What is the difference between merged and short fragments?

Raw YouTube cues break mid-sentence every two or three seconds because they are sized to fit on screen, not to be read. Merged mode rejoins them into sentences and paragraphs, which is what makes a transcript usable for research, quoting or feeding to a model. Short fragments keeps the lines close to what YouTube ships, which is useful when you are checking against the on-screen captions.

Can I use SRT and WebVTT from here in my editor?

Yes, both are standards-compliant and load in Premiere, Resolve, CapCut and any player that reads sidecar subtitles. Because the timings are reconstructed rather than original, treat them as a starting point you nudge into sync rather than a finished subtitle file.

Is there a length limit?

Roughly an hour of speech per video. A longer video is not refused; the transcript stops at about the first hour and says so, which keeps a three-hour podcast useful instead of erroring out.

Free Tool

YouTube Transcript Generator

Paste a link and get the whole caption track as text you can actually read — fragments merged into sentences, ready to copy, download or drop into an editor.

Plain textTimestampsSRT and WebVTTNo signup

Your video

YouTube link

Watch, share and Shorts links all work.

Format

Switching format is instant. The video is only read once.

Line length

Fragmentary cues joined into sentences and paragraphs — readable, quotable, and what you want for research.

Transcript

Paste a link and the full caption track appears here, ready to copy or download. Nothing is uploaded, no account is needed, and the video itself is never downloaded — only its captions, which is why it comes back in seconds.

Raw captions are not a transcript

Every free transcript site dumps YouTube's caption cues at you unchanged, and they are almost unreadable on purpose. Cues are sized to fit two lines on a phone screen for two seconds, so they break mid-clause, repeat words across boundaries, and on a manual track arrive with the line breaks stripped and the words fused together. Paste that into a document and you have text you still have to fix by hand. Merging fragments back into sentences is the entire difference between a transcript you can quote and a wall you have to re-read.

The second thing worth knowing is that a flattened caption track has no reliable timing left in it. Tools that hand you an SRT off the back of one are reconstructing the timings exactly as this page does, and most of them do not tell you. Reconstructed timings are fine for finding the moment someone said a thing; they are not fine as a subtitle file you burn into a video and never check. Anyone promising frame-accurate subtitles from a scraped transcript is selling you a guess with more confidence than it deserves.

Finally, treat auto-captions as a lossy source. YouTube's speech model mangles names, product terms, accents and anything technical, and no amount of formatting downstream recovers what it never heard. For research that goes in front of a client, verify quotes against the video before you use them. For everything else — skimming a two-hour interview, pulling a fact, feeding a model something to work from — it is more than good enough, and it takes ten seconds instead of an hour.

Frequently asked questions

How YouTube transcripts actually work

It reads the caption track — the same text YouTube serves when you switch subtitles on — which is a few tens of kilobytes rather than a few gigabytes. That is why it returns in seconds and why nothing is ever uploaded or stored on your machine.

More free tools

YouTube video summarizerThe same caption track, condensed to a summary and the key points.

Subtitle converterMove between SRT, VTT and plain text, and fix drifting timings.

Script timerHow long a script runs at your delivery speed, before you record it.

AI script generatorTurn what you found in the transcript into a script of your own.

Reading the Transcript is just the tip of the iceberg

Faceless works the other way round: describe a channel and it writes, voices, captions and edits the videos, then posts them on a schedule — captions included from the start rather than scraped back out later.

Script, voiceover and captions generated
Any niche, any format
Posted to YouTube and TikTok for you

Start a channel with AI

Research other people's videos, publish your own.

Ready to Create Your Own Videos?

Start creating amazing AI-powered faceless videos in minutes with Faceless

Start my channel View Pricing