Yes, it says somewhere on the site that it uses soundwaves to share the initial key with your device.
But that's NOT an excuse to make the device depend on the Mobile ecosystem.
If it can play a code on the speaker, it can just talk. In English. Via thirty-six audio files: a.wav, b.wav, ... 9.wav. Ok, thirty-seven recordings--can't forget welcome_to_your_new_router_your_key_is.wav. :)
I'd wager that whatever they're doing, it's far more reliable than reading out English letters and doing speech recognition on it. Accomplishing the key exchange might not have been reliable in all browsers.
But that's NOT an excuse to make the device depend on the Mobile ecosystem.
If it can play a code on the speaker, it can just talk. In English. Via thirty-six audio files: a.wav, b.wav, ... 9.wav. Ok, thirty-seven recordings--can't forget welcome_to_your_new_router_your_key_is.wav. :)