Speechy helps you connect your application entities and their pages to your users voice. It simplifies the recognition workflows, as it focuses you on basic sentence constructs - verbs, nouns and attributes. Using sentence constructs, you can easily lead the user to the wanted page, product or even product list without typing and clicking -- in Minority Report style.
Speechy is a wrapper around Web speech API -- to ease up building speech recognition for your frontend app. Currently, the only browsers which have implemented that feature, are Chrome, along with its Chrome Mobile and Samsung.
Example request:
- Double press the button S on your keyboard
- Say:
Open a user with id 45
Example response:
{
verb: 'open'
noun: 'user',
attribute: {
name: 'id',
type: 'number',
value: '5'
},
isConstructed: true
}
With this response, you can, for example, lead the user to your to page /users/5
In case of a wrong pronunciation or undetected verbs/nouns, you would just get
{
isConstructed: false
}
Setup is easy! Its just 4 easy steps
- Include
speechy.js
as a dependency (whichever way you want) - Set the language, verbs, noun and attributes you want to detect:
var language = 'en-US';
var verbs = [
'open',
'get',
'show',
'give',
'want',
'need'
];
var nouns = [
'user',
'product'
];
var attributes = [
{ name: 'id', type: 'number' },
{ name: 'list' }
];
var cancelPhrase = 'stop';
- Set the required handler for the construct before initializing Speechy
Speechy.onConstructParsed = function (construct){
console.log(construct); // do whatever you like with the received construct
};
- Initialize Speechy with the constructs:
Speechy.init(language, verbs, nouns, attributes, cancelPhrase);
Double press S and voila!