YouTube Captions to Word

Your video: download the caption file from Studio

If the video is on a channel you manage, YouTube Studio is the path that produces a file this converter can turn into timed paragraphs. Sign in at Studio, open Subtitles (or open the video and then Subtitles), and find the language row. The download control sits on that row — hover the row if you do not see the three-dot menu — then choose Download.

Choose .srt or .vtt. YouTube also offers .sbv (SubViewer). SBV times look like 0:01:04.200,0:01:07.800 with a comma between start and end and no --> arrow. This parser keys off the arrow, the same way a normal SRT or WebVTT cue is written. An SBV file will not become a stack of timed paragraphs. Download SRT or VTT instead of converting SBV by hand unless you already have a reason to keep SubViewer.

Studio can list an automatic track and a track you uploaded. Download the one you mean to publish. The automatic track is the one with the recognition mistakes. A track you previously uploaded should come back as you uploaded it, which is what you want if Word is only the reading copy and YouTube already has the captions you checked.

YouTube’s own help documents SRT as plain UTF-8 without style markup. That matches what you should paste here. If a download opens with broken accents, re-save as UTF-8 before converting. Do not pass the file through a word processor’s “save as Web page” on the way.

What Show transcript is — and is not

On a watch page, Show transcript opens a scrolling text panel. You can copy it. That is a reasonable way to grab notes from a video you are watching. It is often not an SRT. The panel’s times are coarse labels next to phrases, and the clipboard text usually has no --> cue arrows and no cue numbers. Pasted into the converter, that text may produce no cues at all, because there is nothing to parse as SubRip or WebVTT.

If the paste preview says no cues were found, you do not have a subtitle file yet. For your own video, go back to Studio and download SRT or VTT. For someone else’s video, use the transcript panel as notes, or ask for the caption file from whoever owns it. This site does not fetch YouTube URLs, and it does not turn a link into captions. The work starts when a file or a real cue list is already on your machine.

WebVTT from Studio is the same job as any other .vtt. Voice tags become speaker labels when they exist; YouTube auto-captions usually have none. Positioning and style settings are dropped. VTT to Word is the format page if the file extension, not YouTube, is the question you actually had.

Auto-captions need a text pass, not a second export

YouTube’s automatic captions share habits with other speech-to-text SRTs. Lines are short so they fit the player. Punctuation is lighter than prose. Music and applause may show up as tags or as guessed words. Names are phonetic. The clocks are usually close enough for a reading transcript and occasionally late or early by a beat.

Leave Merge same speaker on so those short lines join into paragraphs (unlabeled lines merge when the gap is about 2.5 seconds or less). Leave Show timestamps on for the copy you will check against the video, then download again with timestamps off if the published piece should read as prose. The converter will not delete a wrong name. Neither will downloading the same track a second time.

A hallucination-style tail (“thanks for watching” on a video that never says it) is the same cleanup as any other ASR file. Use Whisper SRT to Word if the captions were produced outside YouTube and the artifacts look like model loops rather than YouTube’s line length. Use fix SRT timing only when the words are fine and the clocks walk away from the audio. Do not resync a whole video inside the Word file.

From the .docx to a description, a blog, or Docs

1. Download SRT or VTT from Studio for a video you manage.
2. Skim the first cue and the last cue in a text editor. Delete music-only cues you do not want in prose.
3. Drop the file on the SRT to Word converter. Check the preview: paragraphs, not cue numbers.
4. Download .docx. Open it in Word, Word Online, or upload it to Drive.
5. If the next stop is a shared review, follow SRT to Google Docs so comments land on sentences instead of on subtitle codes.

Chapters and descriptions are not caption files. Pasting a chapter list will not create timed transcript paragraphs. If you also have CapCut or Premiere captions for the same video, prefer the file you have already corrected. A Studio auto-track and a CapCut auto-track of the same talking-head video will not match line for line; pick one source and edit that. CapCut’s export dialog is covered in CapCut subtitles to Word.

Privacy and other people’s videos

Conversion stays in the browser. Studio is where the official caption download lives, and it is limited to videos you can manage. Do not paste a caption file you are not allowed to republish into a public Doc and call it your transcript. The tool will still format it. Permission is your problem, not the formatter’s.

For a meeting or a webinar that was never on YouTube, you do not need this page. A Zoom cloud transcript is already VTT or a speaker-and-timecode paste, and it has a dedicated path at Zoom transcript to Word.

Related workflows

• VTT to Word
• Whisper SRT to Word
• SRT to Google Docs
• All subtitle-to-Word guides
• SRT to Word converter — same tool the buttons above open, with these toggles already set.