BT

Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ

Topics

Choose your language

InfoQ Homepage News Amazon Adds Finer Granularity of Control to Their Voice Recognition API

Amazon Adds Finer Granularity of Control to Their Voice Recognition API

This item in japanese

Bookmarks

Amazon's Alexa Voice Service API, the NLP (natural language processing) API that powers Amazon Echo, has a new update that allows for developers to use Alexa to turn any device into a "smart" device through the use of the API's voice recognition features.

The Alexa API is one of the latest of several large scale NLP releases by tech giants, such as Microsoft and Google. Alexa, in particular, specializes in the integration between devices and voice recognition, much like IBM’s recent work with APIs on their cloud platform Bluemix. Alexa is designed to be a developer API for companies looking to fine tune the growing market of "internet of things" (IOT) devices. A key issue with current smart devices is they all exist on different platforms, with varying levels of capability and security. A home's Hue light does not know how to talk to its August smart lock, creating a barrier to more advanced and usable IOT phrases, such as “I’m leaving the house”, which would create an automatic chain of commands that lead to turning off the lights and locking the doors. Instead you are left to manually fiddle with a plethora of apps, each with a different user interface and login. Apple’s HomeKit has provided a lot of progress in this area, but little in terms of customizable voice recognition APIs for development, HomeKit integrates directly with Siri.

A key feature of the Alexa API update is that it provides finer granularity around voice activated media, such as volume and playback control, as well as better integration with external devices. You do not have to have an Amazon device to use the API and can register any Raspberry Pi or other custom hardware with it online. This includes support for mobile devices and server-side messaging. This means, you can design your own voice user interface, built with a custom programmed set of skills on your own hardware. Many NLP platforms are entirely software focused like Google's Cloud Natural Language API, or provide a large amount of device integration support but sacrifice on customizability that may jeopardize standards set in an all encompassing eco-system of applications, like Apple's HomeKit. Alexa is unique in their API additions, by offering a focus in voice recognition capabilities while not pressuring the user into buying into a total platform solution that may come at the cost of flexibility.

Some of the companies jumping on the Alexa bandwagon are Petnet, a voice activated pet feeder and Mojio, an in-car application control system. These applications can now take advantage of voice activated timers and alarms, integration with the Alexa mobile applicaiton, and new features around system level control, like "mute" and "unmute" capabilities. Alexa has also introduced a new collaboration with Intel to release what they call Form Factor Reference designs, new Intel board designs that aim to to cut down on development time and help device manufactures further create smart devices through hardware that is designed to pair with their API software.

Rate this Article

Adoption
Style

BT