Swype has launched the "next generation" of its predictive keyboard, promising four methods of text entry including voice recognition and handwriting. The Nuance-powered system supports 55 languages and automatically learns from users' speech and typing to improve its recognition and prediction.
Swype uses the same historical learning system as Nuance's XT9 platform, and can be given an extra glimpse into your writing style by sifting through emails and other messages already on your device. They're mapped into Swype's unified language model, cutting out the tedious process of manually entering acronyms, technical terms or unusual names.
The process is copied into both text and voice user-dictionaries, which means voice recognition is improved by written examples too. There are four different keyboard options: swyping from letter to letter; typing rapidly using predictive XT9 text; speaking to their device using the voice recognition system; or using their fingertip to draw out words, letters and symbols. Each of the four can be switched between on the fly.
Swype is already offering the new version of the keyboard to its device and carrier partners with Android phones and tablets, and has an SDK for other platforms as well. Individual users will be able to try it out by registering at the Swype beta portal.