SRT or VTT: which subtitle format should you use?

The short answer: use WebVTT if the video plays in a browser, and SRT almost everywhere else. Both are plain text files, and the differences between them are small but strict.

The same cue in both formats

1
00:00:04,000 --> 00:00:09,200
Before we start, this recording
stays on my phone.

WEBVTT

00:00:04.000 --> 00:00:09.200 line:0
Before we start, this recording
stays on my phone.

According to MDN, the WEBVTT line is the only required part of a WebVTT file, and the file must be UTF-8. Timings can be written as mm:ss.ttt or hh:mm:ss.ttt, and the end time has to come after the start time. SRT numbers its cues and uses a comma before the milliseconds.

What WebVTT adds

SRT is close to bare text and timings. WebVTT was built for the web, so it can do more:

There are limits. MDN notes that STYLE blocks must come before any cues, and that they style appearance rather than position. Characters such as & and < need escaping in cue text, a detail that’s easy to miss if you edit files by hand.

Which one to choose

• Importing into a video editor or uploading to a platform? Start with SRT, the most widely supported format. Check the tool’s help page if you’re unsure.
• Playing video on your own site? Use WebVTT in a <track> element.
• Need captions placed around on-screen text? Use WebVTT and its cue settings.
• Not sure? Keep both. It costs nothing, and you’ll have whichever one the next tool wants.

Automatic captions still need a person

Captions aren’t only for people watching with the sound off. The W3C describes them as a text version of the speech and non-speech audio needed to understand a video, used by people who are Deaf and others who can’t hear the audio. In WCAG, captions on pre-recorded video with audio are required at Level A.

The W3C is also clear that automatic captions don’t meet those needs unless they are confirmed to be fully accurate. So treat any automatic transcript as a first draft. Read it against the audio, fix names and technical terms, and add important sounds, such as [door slams] or [music stops], that a speech model may leave out.

Making both on your phone

EchoLocal Pro exports SRT and VTT from the same transcript, and the work happens on your phone. Words you don’t change keep their timestamps as you edit, and tapping a word plays that moment, which makes the review pass quicker. You can see both formats in the demo on our home page.