Common to all such services is a need for some form of speech recognition, as well as an engine that can generate speech from digitised text.

ECONOMIST: REPORTS: Just talk to me | The

The multimodal and multilingual speech engine runs on a cloud-based server to provide extremely high-quality recognition with low latency.

ENGADGET: QNX builds in-car speech framework with AT&T's Watson, knows our true intentions

Though it transcribes all of the speech pulled out of videos from its content partners, it only runs other videos through speech-recognition software when there isn't enough metadata with the files for the search engine to figure out what it contains.

FORBES: co-founder Sergey Brin