Free tool

YouTube Transcript Cleaner

Paste a transcript copied from YouTube, or an SRT, VTT or SBV caption file, and get readable paragraphs: no timestamps, no [Music] tags, no broken lines.

How do I clean up a YouTube transcript?

Open the video's description, choose Show transcript and copy the whole panel. Paste it into the box above. Timestamps, [Music] tags and >> speaker marks are removed straight away, and the broken caption lines are joined into readable paragraphs you can copy or download as a .txt file.

How to use the transcript cleaner

  1. Copy the transcript. On YouTube, open the description, choose Show transcript, then select the whole panel and copy it. The guide to copying a YouTube transcript covers each step, and there's help if the transcript button is missing.
  2. Paste it above. The clean version appears as you paste, with a word count and how long it takes to read.
  3. Adjust and save. Untick an option to keep that part, then press Copy or Download .txt.

No transcript to hand? Press Try an example to see a short sample cleaned. On a phone, copying works a little differently; see getting a transcript in the YouTube app.

What it removes, and what it keeps

A copied transcript puts a timestamp on every other line and splits sentences wherever the captions happened to break. The cleaner strips that clutter and puts the words back together, in the same order, without changing any of them.

A transcript copied from YouTube, with a timestamp on every other line, [Music] and [Applause] tags and >> speaker marks struck through, becomes one readable paragraph: 21 lines turn into one paragraph of 59 words.
The sample from Try an example: 21 copied lines become one paragraph, with every word kept.
What each option does
OptionRemoves
TimestampsLines like 1:02:05, times at the start of a line, SRT cue numbers and timing lines such as 00:00:01,000 --> …
Tags[Music], [Applause], (laughter), >> speaker marks and inline caption tags such as <c>
ParagraphsThe line breaks between caption fragments, regrouping the text into paragraphs

Cleaning SRT, VTT and SBV caption files

Caption files hold the same words as the transcript panel, wrapped in numbering and timing. Open the file in any text editor, copy everything and paste it in. The cleaner recognises all three formats. It removes cue numbers, time ranges, the WEBVTT header and styling tags, and turns escaped characters such as &gt; back into what was said. If it's your own video, YouTube Studio's Subtitles page lets you download a caption track as a file. The guide to SRT vs VTT files explains how the two main formats differ.

An SRT file and a VTT file with each line labelled: cue numbers, time ranges, the WEBVTT header and Kind line are removed; lines with [Music], inline c tags or escaped >> marks are cleaned; plain text is kept. Both become: hi everyone and welcome back to the channel.
Different wrappers, same words: both files clean to one identical line.

How the paragraphs are formed

When the captions have punctuation, usually because a person wrote them, the cleaner starts a new paragraph after every five sentences. Auto-generated captions usually have no punctuation at all, so there are no sentences to count. In that case it breaks the text about every 90 words instead, which keeps a long transcript scannable.

Two rows of blocks: punctuated captions split into paragraphs of five sentences, five sentences and the rest; unpunctuated auto-captions split into paragraphs of 90 words, 90 words and the rest.
Five sentences per paragraph when there's punctuation; about 90 words when there isn't.

Who uses a transcript cleaner

  • Students turning a lecture into text they can highlight and quote, as the first step of turning a video into notes.
  • Writers and creators starting a blog post from their own video; the guide to repurposing a video into a blog post picks up where this leaves off.
  • Researchers and journalists who need the exact words for a quote, rather than a paraphrase.
  • Anyone pasting into ChatGPT or another chatbot. Clean text uses fewer tokens and is less likely to confuse it; see summarizing a YouTube video with ChatGPT for the trade-offs.

Key features

  • Instant: cleans as you paste, with no button to press and no waiting.
  • Every common format: YouTube's transcript panel, SRT, VTT and SBV files.
  • Your words, untouched: it removes clutter and never rewrites, summarises or reorders.
  • Choose what goes: keep timestamps or tags by unticking an option.
  • Word count and reading time for the cleaned text. The video reading time calculator compares that with watching.
  • Private: runs entirely in your browser. Nothing is uploaded or stored, and it works offline once the page has loaded.

Examples

What you paste, and what you get
You pasteOptionsYou get
An auto-captioned 40-minute lecture from the transcript panelAll onParagraphs of about 90 words, ready to annotate
An SRT file with human captionsAll onParagraphs of five sentences each
A transcript you want to quote with timesTimestamps and paragraphs untickedEach time kept on its own line above the words, tags removed
A VTT file for a subtitle editParagraphs offOne clean caption per line, no timing

What cleaning can't fix

Auto-generated captions misspell names and technical terms and mishear words, and the cleaner doesn't correct them: it never changes the words. How much that matters depends on the video; the guide to how accurate YouTube's auto-captions are shows where the errors cluster. It also can't separate speakers beyond removing >> marks, or add punctuation that was never there.

Tips

  • Copy the whole panel. Click in the transcript and select to the end; a partial copy cleans fine but leaves gaps.
  • Keep timestamps when you'll cite. Untick Remove timestamps if you need to point back to the exact moment later.
  • Prefer human captions. If a video offers both, choose the uploaded track over "auto-generated": you'll get punctuation and better paragraphs. Captions, subtitles and transcripts explains the difference.
  • Want the argument, not every word? A cleaned transcript is still a transcript. A summary is a different thing and is usually the faster read.

Frequently asked questions

How do I remove timestamps from a YouTube transcript?

Paste the transcript into the box above with Remove timestamps ticked. Every timestamp line, time at the start of a line and caption timing line is removed, and the text is joined back into paragraphs.

Does it work with SRT, VTT and SBV caption files?

Yes. Open the file in any text editor, copy everything and paste it in. Cue numbers, time ranges, the WEBVTT header and styling tags are removed.

Is there a length limit?

No set limit. Transcripts several hours long clean instantly on a normal computer, because nothing is uploaded.

Can it fix the spelling in auto-captions?

No. It only removes clutter and rejoins lines; it never changes the words, so what you get is still exactly what the captions said.

Is my transcript uploaded anywhere?

No. The cleaning happens in your browser, on this page. The text isn't sent to us or stored, so it's safe for private recordings too.

Rather read the point than the transcript?

The summarizer reads the transcript for you and returns the argument, the takeaways and chapters.

Summarize a video — free