Ah, I see. Memrise doesn’t currently have any functionality for creating audio through a computer voice, although I think there may be some userscripts floating around that do that.
However, when you’re entering a word into your database, there should be a dropdown that auto-populates with definitions that other people have used in different courses. If you select one of those and that definition has audio attached to it, then your word will have audio attached to it, too.
If you want to put in the effort I’d suggest you take a look at forvo.com. Here you find rather decent soundfiles for a lot of vocabulary by native speakers. You still need to upload the sound to your own course and add an audio column to do that.
See here: How to put sound in your own course?