SRT vs VTT: which caption file to use where

SRT and VTT side by side, line by line

WebVTT was built on SRT, and W3C's own WebVTT specification thanks the SubRip community for the format it started from. So the two files look almost the same: a number or name, a start and end time joined by an arrow, the caption text, a blank line. SRT never got a formal standard; the Library of Congress calls it a very simple text format that is not well standardized. Here are the same two captions in each.

1
00:00:04,200 --> 00:00:07,000
MARIA: Welcome to the shop.

2
00:00:07,400 --> 00:00:09,800
[door bell rings]

WEBVTT

1
00:00:04.200 --> 00:00:07.000
<v Maria>MARIA: Welcome to the shop.

2
00:00:07.400 --> 00:00:09.800 line:10%
[door bell rings]

Four things change. The WebVTT file opens with the word WEBVTT and a blank line. Its times use a full stop before the milliseconds where SRT uses a comma. A WebVTT time line can carry settings after it: the line:10% above moves the second caption near the top of the picture, out of the way of anything shown low on screen. And the <v Maria> voice tag tells software who is speaking; the name written in the text is what a viewer reads, whatever the player does with the tag.

How to convert SRT to VTT, and back

Renaming clip.srt to clip.vtt is not a conversion: the file still has no WEBVTT line and still has commas in its times, so a web player that reads WebVTT will not accept it. Most caption editors export both formats, and our closed captioning software comparison lists which ones. By hand in a text editor, SRT to WebVTT takes four steps.

1. Type WEBVTT on a new first line and leave one blank line under it.
2. Replace the comma in every time with a full stop: 00:00:04,200 becomes 00:00:04.200. Change only the commas inside times, never the ones in the text.
3. Keep the caption numbers or delete them. WebVTT treats them as names, so each must appear only once.
4. Save the file as UTF-8 with the .vtt ending.

WebVTT to SRT runs the other way: delete the WEBVTT line and any STYLE or NOTE blocks, cut the settings off the end of each time line, put a comma back before the milliseconds, write any time without hours in full (00:04.200 becomes 00:00:04,200), and number the captions 1, 2, 3. Whatever positioning or styling the WebVTT file had is lost.

Which one WCAG asks for

Neither. 1.2.2 Captions (Prerecorded) is a Level A criterion in WCAG 2.1 and 2.2, and it asks for captions for all prerecorded audio in synchronized media. It names no file format. What it judges is what the captions say, so an accurate SRT meets it and a careless WebVTT file does not.

The format matters once the video plays on your own page. W3C's technique H95 shows captions added to an HTML video through a track element like <track kind="captions" src="myvideo_en.vtt">, and the track element reads WebVTT files. Set the kind to captions, not subtitles: subtitles carry only the dialogue. For everything else captions must carry, read our video caption requirements.

A file a platform made for you is a draft. W3C says automatically generated captions do not meet user needs or accessibility requirements unless they are confirmed to be fully accurate; they can be a starting point. Download the draft, fix names, numbers and sound tags, then upload the corrected file. If you would rather pay a person to do it, compare captioning services.