Vlingo Gets $6.5 Million Funding For Speech-To-Text For Mobile Web Browsing


Vlingo, a Cambridge, Mass-based start-up, has launched a beta version (from its website) of its voice-powered interface for mobile phones, which will let people talk to enter text into any vlingo-enabled text box on the handset. So it’s not limited to the vlingo app but is intended to be used with all the mobile phone text boxes, and will be particularly useful with web surfing. The promise is that people won’t have to use specific commands or change how they speak, they just say whatever they want and vlingo inputs the text word-for-word. I’m naturally skeptical about this, having had too many speech-to-text companies over-promise and under-deliver, but if they can get it right it should be pretty useful. The vlingo API can run on the majority of 3G and multi-media enabled phones. The service lets people change between keypad entry and speech entry and learns from its history — that is, it takes into account what each user has spoken in a particular text box within a particular application and uses this to improve accuracy for that user and the entire community. While this improves general accuracy, VentureBeat has some comments from CEO Dave Grannan admitting that “the downside to this approach in the short term is that regional accents or foreign words may not get understood the first time…Because the technology detects details of vocabulary and speech patterns, internationalization to other accents and languages is a complex task, and a ways down the road”. Vlingo has received $6.5 million in first round funding from Charles River Ventures and Sigma Ventures.
The demo is here, and looks pretty good.

Comments are closed.