VTT to Text Converter turns a WebVTT caption file into readable text or a Word document. WebVTT is the caption format of HTML5 video, and it is what you get when you download captions from YouTube, Vimeo, Zoom, Microsoft Teams recordings, Loom and most online course platforms. A .vtt file holds much more than the words: a WEBVTT header, NOTE and STYLE blocks, cue settings such as align:start position:0% after every timestamp, inline word timings, and voice tags like <v Anna> that name the speaker. This page removes all of that and keeps the speech, joined into sentences and paragraphs. Speaker names from voice tags are kept, and every change of speaker starts a new paragraph, so a Zoom or Teams meeting transcript reads like a conversation, Anna: ... Ben: ..., instead of a list of captions. YouTube auto-generated VTT files repeat each line two or three times as the captions roll up; those copies are removed so each sentence appears once. Download the result as .txt, or as .docx with speaker names in bold, which is the quickest way to get VTT into Word. SRT, SBV and ASS files work here too, since the format is detected automatically. Nothing is uploaded. To transcribe a recording that has no captions yet, use Speech to Text Audio Transcriber, and for SRT files see SRT to Text Converter.
An SRT file is a plain text file made of numbered blocks called cues. Each cue has a sequence number, a timing line such as 00:01:04,250 --> 00:01:06,900, and one or two lines of text that are kept short on purpose: subtitle guidelines from broadcasters such as the BBC and Netflix limit a line to about 37 to 42 characters and a cue to two lines, so viewers can read it in the few seconds it is on screen. That is why simply deleting the numbers and times leaves a column of fragments: each sentence is spread over two or three cues and four or five lines. WebVTT, the W3C format used by HTML5 video and by YouTube, works the same way but starts with a WEBVTT header, uses a period before the milliseconds and can carry styling, positioning and voice tags that name the speaker. YouTube SBV is an older, simpler format with the start and end time on one line, and ASS/SSA, popular for anime fansubs, stores each line as a Dialogue entry with a style and the name of the character. To rebuild a transcript, this tool first reads every cue with its start and end time, strips markup and sound descriptions, and then joins the text into one stream of words. The timing decides the paragraphs: the gap between the end of one cue and the start of the next is a reliable sign of a pause, and a gap of two seconds or more usually means a new thought or a new scene. Auto-generated YouTube captions need extra care, because they roll up: each cue repeats the previous line above the new words, and between them sit cues only 10 milliseconds long. Read naively they print every sentence two or three times. The converter recognises this pattern and keeps each line once, while ordinary subtitles where someone really says the same thing twice are left as they are.