Convert a YouTube video to text

Why the output reads properly

YouTube stores captions cut into fragments a few seconds long, because that's what fits on screen. Dumped straight out, that gives you something like this:

and that's cool
the cool thing about these
guys is that they have really
really really long trunks

Technically a transcript. Useless to read. We reassemble those fragments into paragraphs using sentence endings and the natural pauses in speech, so you get prose instead of confetti. The words are untouched — only the line breaks change.

For auto-generated captions, which carry almost no punctuation, we lean on the timing gaps instead. It's not magic, but it turns an unreadable block into something you can skim.

Three views, depending on the job

• Paragraphs — for reading, quoting and pasting into a document.
• Timestamps — every line timecoded and clickable, for finding a moment.
• Plain — one continuous block, for feeding into something else.

What people use this for

• Studying — turning a lecture into notes you can search and annotate.
• Research and journalism — pulling accurate quotes with a timecode to check against.
• Reading instead of watching — a forty-minute video is about five minutes of reading.
• Accessibility — text that works with a screen reader or can be enlarged.
• Translation drafts — many videos carry caption tracks in dozens of languages already.

Getting it into your notes

The Copy button copies whichever view you're looking at, so copying from Paragraphs gives clean prose with no timecodes. If you'd rather have a file with the video title and a link back to the source already in it, download the Markdownexport — it drops straight into Obsidian, Notion or anything else that reads Markdown.

Common questions