It sounds like [θræʊt], doesn't it? What I did next was to start with the first 30 ms of the word and then add 30 ms each time till the end of the word (30, 60, 90, ... 300, 330, 360 ms). Listen to this:
Here's the first word followed by the next one. Can you decode what she wants to say?
The problem of understanding this phrase (and many others in her speech) lies in the fact that she compresses a disyllabic word to a monosyllable. What may make things even more complicated is the fact that this compression happens at the very beginning of her speech when the listener's ears and brain have not become attuned to her dropping-one's-syllables style.