AT&T brings out HTML5-ready Watson API
Telecoms giant AT&T has officially released the free application programming interface (API) for its Watson-powered speech program which will allow users to access their apps via voice control.
The intuitive text-to-speech software development kit (SDK) will be available for both native and HTML5 apps and is designed to wade through any accent or dialect to recognise what is being said – a previously complex problem.
Among the solutions the API will aim to provide are:
- A voicemail-to-text service which is trained on a “massive” amount of data from call centres
- An SMS service which can instantly bring a spoken message through as text
- A search facility which incorporates everything from web to business and TV queries
- A “babelfish” service which generically transcribes English and Spanish languages
Perhaps the most intriguing concept of the software is that each of these tasks are put into their own individual context – there are seven in total – which captures better meaning and gives more focused results. An example of this is the TV context, which ties into AT&T’s U-Verse TV guide.
The API is free – until the end of this year at least – for any developer who has paid the $99 yearly registration fee. From 2013, a points-based transactional system will be brought in.
AT&T had said they were planning on opening Watson up to developers as far back as April, so they have made good on their promise.
John Donovan, senior executive VP of AT&T technology and network operations, said of the release: “It’s a technology that’s been a long time in development and more than 600 patents in the making, and we’re excited to open it up to developers and see what they make of it.
“Our goal is to make it as easy as possible for developers to integrate AT&T Watson speech recognition technology into their apps, so we encourage developers to take advantage of our Speech SDKs”, he added.
Elementary, my dear...?
Benoît Felten had recently written about how, back in the early 1990s, AT&T had foreseen a variety of trends, including video conversation, online ticketing and e-reading, but failed to capitalise on the developments at Bell Labs.
The article brought about a fierce riposte from those who were there at the time, with one commenter claiming the government was out to take AT&T down and stating that “so many times during the 80s and 90s T initiated efforts to move forward but was blocked”.
So is AT&T Watson going to be a big hitter and maybe realise some of AT&T’s potential?
The main competition appears to be Apple’s Siri, although it appears that Watson will just provide a voice-to-text solution for app developers, rather than being an overall personal assistant like Siri. Same goes for the new Android alternative, Robin.
As a result, it would appear that AT&T Watson would move into the realm of products such as Nuance instead.
What do you make of this API and SDK? Would it be beneficial for you as a developer?