Hi Alex,

Yes, that's what is in their documentation and I tried that initially but the config section threw errors when I tried to use it. After looking at the jsfiddle code that the documentation has a link to, and emailing Talkify themselves, I ended up with the following working code:

<script type="text/javascript">

    talkify.config.remoteService.host = 'https://talkify.net';
    talkify.config.remoteService.apiKey = 'my-key';


    talkify.config.ui.audioControls = {
      enabled: false, //<-- Disable to get the browser built in audio controls
      container: document.getElementById("player-and-voices")
    };

    document.getElementById("talkifyButton").onclick = function () {
        var player = new talkify.TtsPlayer();

        new talkify.playlist()
            .begin()
            .usingPlayer(player)
            .withTextInteraction()
            .withElements(document.querySelectorAll(".readMe")) //<--Any element you'd like. Leave blank to let Talkify make a good guess
            .withRootSelector('body')
            .build() //<-- Returns an instance.
            .play();
    }
</script>

I took out the text highlighting because that stripped out the styling from whichever piece of text was being read. So that I could decide exactly which bits of the page to read I added a .readMe class to the relevant elements and used that in the withElements parameter.

Hope that's useful to anyone else trying to do this.

Cheers, Alistair