This is a demonstration of using Api.ai and the Web Speech APIs (via the excellent SpeechKITT library) to add a voice interface to a web application. In this case, that application is the Spree e-commerce platform.

In order to get this up and running, you'll need an account at Api.ai. Create a new agent, and upload the entities and intents found in app/assets/voice. Then create a .env file and add the following line:

API_AI_CLIENT_ACCESS_TOKEN=

Be sure to add your agent's client access token to the right of the =, obvs.

From there, it should just be the normal Rails app setup rigamarole.

Check out the diff between the demo and master branches to see what was done.