This repository is an example, developed in classroom, for getting started with voice user interfaces.
It uses the Web Speech API (experimental) for speech-to-text and text-to-speech operations. For Natural Language Processing, it uses Dialogflow agent, included as a .zip file. Composer is used to manage the dependencies, i.e., the Dialogflow PHP client library.
Tested in Chrome 57+, Firefox 70+, and Safari 10.1+. Please, notice that speech-to-text capability does not work in Safari and Firefox.
To get started with the project, import the WeatherAgent.zip into Dialogflow, generate a service key from Google (in JSON), and download it into the root of the project as client-secret.json.