This project parses to JSON and serializes back to the .vtt format, not to HTML. This makes it easy to manipulate the cue timeline (e.g. adding time between two cues).
It does not follow the [http://dev.w3.org/html5/webvtt/#parsing](WebVTT parser algorithm) and is not meant as an implementation of WebVTT. For that, you should look to the node-vtt project. It also does not validate .vtt files. For that, you should look to the webvtt project.
var fs = require('fs'),
vtt = require('vtt-json');
var input = fs.createReadStream('input.vtt');
var output = fs.createWriteStream('output.json');
input.pipe(vtt.parse())
.pipe(through2.obj(function(obj, enc, cb) {
cb(null, JSON.stringify(obj, null, '\t'))
}))
.pipe(output);
Files can also be converted from .vtt to JSON on the command line.
vtt-json parse input.vtt --output=output.json