Инструменты для соцсетей IT Market
Удалите номера и таймкоды из SRT/VTT-субтитров и получите чистый текст для статьи, описания или перевода.
Tool guide
Subtitles pulled off YouTube arrive as a ladder of cue numbers, «00:00:01,000 --> 00:00:03,000» lines and half-sentences. Nobody can read that, and a translator or language model fed with it starts translating the timecodes themselves. Paste the SRT or VTT content into the subtitle box and press «Generate» — only the spoken words remain. It runs in the browser with no upload, and the output is ready for a transcript, a video description, an article or translation. See also: build subtitles back from plain text, clean an auto-generated transcript, shape the text into a post caption.
Yes. VTT differs from SRT by using a dot before the milliseconds, a WEBVTT header on the first line, and optional cue settings such as align:start or position. All of that is recognised and stripped along with the cue numbers, so a subtitle export from YouTube Studio can be pasted in unedited.
Because the subtitles did not have any. YouTube's automatic captions rarely include punctuation and break phrases by line rather than by meaning. The tool removes the technical lines and does not invent punctuation on your behalf. Adding full stops and paragraph breaks is a manual pass, or a job for a language model afterwards.
No, repeats stay. YouTube auto-captions often overlap, repeating the last line of one cue as the first line of the next. This tool takes out numbers and timecodes only; duplicated phrases have to be cleaned in a text editor with find and replace, or caught by reading through before you publish.
No practical one — parsing happens line by line in the tab's memory. A two-hour film's subtitles run to roughly a hundred thousand characters and go through without noticeable delay. The only slowdown is on a very old phone, where pasting that much text into the field is itself the slow part.
Nowhere. Parsing runs in the page script inside your own tab, no network request carries the text, and nothing survives closing the page. That makes it safe for the transcript of a private call or an interview that has not been published yet.
The reverse trip runs through the companion text-to-SRT tool, which re-splits the text into cues with evenly spaced timings. The original timings are gone for good once you strip them, so the new ones are approximate. If you only need to reword lines while keeping the cue grid, edit inside the SRT file instead.
Tags such as and in SRT, or markup like in YouTube auto-captions, stay in the text when they sit inside a spoken line. Remove them with find and replace in a text editor — usually a single pattern does it, since the same tag repeats throughout the file.
Before: 1 00:00:01,000 --> 00:00:03,000 Hi everyone 2 00:00:03,000 --> 00:00:06,200 today we are talking about subtitles
After: Hi everyone today we are talking about subtitles
Before: WEBVTT 00:00:02.400 --> 00:00:05.100 align:start this is the intro to the lesson
After: this is the intro to the lesson
Your rating and feedback help decide what to improve next.
