In this Reply we correct a mistake that we made in the correctness proofs of our protocol. Specifically, the Bell inequality we used ensures security but does not allow us to produce a key. In this Reply we explain and correct this mistake by adjusting the Bell inequality we used in the proof. Incidentally, this correction leads to slightly better asymptotic key rates. Importantly, none of the conclusions of the article are affected.