It seems to me the easiest way to do this would be with a Raspberry Pi and one of the wake-word detection engines used in voice assistants. Some quick Google searching seems to indicate that Snowboy allows for user defined wake-words. So you could train it to pick up on "Frau". After that, it would just be a script to play an MP3 of the horses.
I need help figuring out how to build this. Any suggestions? I'm not an electrical engineer so any help would be appreciated.