Yandex has launched a new speech recognition cloud service called SpeechKit Cloud. It is designed for developers who want to make their products respond to human voices.
The company says that integrating the Yandex technology is very simple. All you need to do is read a few pages of documentation and write a couple of lines of code into your software.
The feature can be built into anything at all, from computer games to satellite navigation systems for vehicles. Commands are processed on Yandex’s servers, but according to the company recognition only takes about a second.
SpeechKit Cloud works with both Russian and Turkish. The recognition statistics for Russian are: dictated text – 82%, short queries from 3-5 words on any subject – 85% and location dependent queries – 95%.
Several companies took part in the testing of SpeechKit, including Oktell (call centre software) and Cubic Robotics (developers of domestic robot Cubic). For Oktell the system was used to recognise caller’s voices and connect them to the correct specialist, and the robot Cubic used it to listen and respond to questions.
Access to the service is now open to all developers. SpeechKit Cloud is free for the first month, after which the cost depends on the demands of the user.
SpeechKit Cloud is based on the speech recognition technology Yandex SpeechKit, which they launched in 2013, and is now used by 400 mobile apps for Android, iOS and Windows Phone and processes millions of speech queries every day.
Source: RusBase.vc, CNews.ru
