Loading page
About this tool
Turn YouTube caption text or an .srt, .vtt or .json3 caption file into a clean, paragraphed transcript with timestamps, word count and reading time.
The YouTube Transcript Generator converts YouTube caption data into a clean, readable transcript: paste the text from a video's "Show transcript" panel, or drop in a downloaded .srt, .vtt or .json3 caption file, and it rebuilds the words into paragraphs with timestamps. It parses all four caption formats, strips [Music] style sound labels and the repeated lines that YouTube's rolling auto-captions produce, then reports word count, caption duration, speaking rate and reading time at 238 words per minute. It is built for creators writing show notes, students turning lectures into study notes, and marketers repurposing video into blog posts — all in the browser, with nothing uploaded to a server.
Open YouTube Transcript Generator on AltFTool — it loads instantly in your browser.
Upload or drag in the file you want to work with.
Adjust the settings and preview the result.
Download the finished file to your device.
Reads SubRip .srt, WebVTT .vtt, YouTube json3 and plain transcript-panel text without any setup.
Rolling auto-caption cues that repeat the previous line are de-duplicated, so the word count is honest.
Caption text is parsed on your device and never uploaded, so private or unlisted video captions stay with you.
Open the video, click the three-dot menu under it (or the "...more" area in the description) and choose "Show transcript". The panel opens beside the video; select all of it, copy, and paste it here. This tool cannot fetch captions from a URL by itself, because YouTube does not allow a web page in your browser to read another site's caption data.
YouTube's automatic captions roll: each new cue repeats the line before it so the text stays on screen. This tool removes that repeated prefix, so a cue reading "hello there" followed by "hello there friends" counts four words, not five. Switch the de-duplicate option off if you want the raw cue text preserved exactly.
Word count divided by 238 words per minute, the mean silent reading rate for English non-fiction from Brysbaert's 2019 meta-analysis of reading-speed studies. Speaking rate is a separate figure: transcript words divided by the caption span in minutes, which for most presenters lands between 120 and 160 words per minute.
A gap in speech of 2 seconds or more between caption cues starts a new paragraph, and a paragraph is also closed once it reaches 90 words. You can move the gap anywhere from 0.5 to 10 seconds — a smaller gap gives short, dialogue-like paragraphs, a larger gap gives long prose blocks.