Voicery synthesizes the most realistic human voices using deep neural networks. Prior to starting Voicery, Andrew, one of the founders, led the speech synthesis research team at Baidu Research. By synthesizing speech nearly indistinguishable from human, Voicery aims to enable new media applications, such as auto-generated audiobooks, podcasts, television dubs, and voice overs. Their website lets users compare human and synthesized sentences and guess which is which.