Wow, that was such a cool experience! At first I was a bit overwhelmed , but after a short ajusting period it was really fun to mentally map out the level based on the audio feedback. Maybe some audio indicators about the state of the battle would also be cool (like how far the enemies are away from the core) would be cool (I hope I haven't missed it), since I was mostly relying on visuals for that.
There seems an encoding issue with some of the text: on my system, the text-to-speach reads " ' " characters as "Euro trademark", so "I'm" became "I Euro trademark m" which made following the narration a bit difficult :D
Anyway, awesome work!