Amazon's Alexa Voice Service let use voice to integrate with 3rd party products, by way of developing Alexa Skills.

Software

AlexaPi is a Python client for Amazon's Alexa service. This project, ALL does a nice job at integrating Alexa, Lambda, an RTOS RPi distro, and describing a bit of machine learning in the way.

Hardware

EasyVR has an Arduino shield that provides voice recognition, which you can find at sparkfun.

For raspberry pi, there's the Matrix which has an 8x MEMs microphone array with Alexa support. You can buy them here from $100.

There are also multichannel USB microphone array like the UMA-8 that goes for $95, which has learning resources here

Voice as Password

Currently- Feb 2017- AWS does not support voice biometrics authentication or forwarding voice input to skills. In fact, the way it works is all voice input gets translated to text, which is the data sent to the AWS service.

Cloud service VoiceIt provides cloud based voice biometrics services.

There are other services, like armorvox which let you try out. Or like vocalpassword that are more enterprise focused.

Speaker recognition would be needed in order to have independent Alexa's placed around and the member just access any.

The HSBC UK rolled out a service for banking as well as Australian Tax Office.

Microsoft has a cloud service for voice recognition, Speaker Recognition API.

Speaker Recognition API

Select a pass phrase from the given list of phrases. Use that phrase and record three audio samples to register your voice with the service, this step is called "enrollment". After your enrollment is completed, you can start the verification step using a different voice recording or phrase to test the service.

Resources

PocketSphinx is a speech recognition engine that can be used for hotword detection.

Node: