Mian Asad Ali

Get started

Spoken is a free SDK for voice controlled apps. Improve your user's experience with easy to use Human Interface.

Step 1Import Spoken
You can import the spoken module in your app easily using the following example.


              
              npm i spoken


              
import spoken from './node_modules/spoken/build/spoken.js';

    spoken.say('Should I turn the hallway light on?').then( speech => {

    spoken.listen().then( transcript =>

        console.log("Answer: " + transcript) )

} )


              
  spoken.say('Should I turn the hallway light on?').then( speech => {

    spoken.listen().then( transcript =>

        console.log("Answer: " + transcript) )

} )

Step 2Text-to-Speech
Synthetic voices are pretty good these days. You can still tell they are robot voices.
Turn text into real-world spoken voice. The voice is synthetic. You can pick from a few different voices too.

             

// Hello World

spoken.say('Hello World.');
  
// Say a quick message as Guy Fieri

spoken.say( 'Looks like your on a trip to flavor town.', 'Daniel' );
              
// Speak with Damayanti's voice

spoken.say( 'I am princess of the Vidarbha Kingdom.', 'Damayanti' );

Step 3Promises and Await
This SDK Supports modern async programming.

              

// Promise

spoken.say('Hello World.').then( v => console.log('Done talking.') );
            
// Async/Await

await spoken.say('Hello World.');

console.log('Done talking.');

Step 4Voice Selection
Synthetic voices are selectable using the spoken.voices method. There is a default voice that is included. You may want to pick a different voice. The following examples show you how to select a differnt voice.
Get list of English voices.


              
// List of English Speaking Voices

spoken.recognition.language = navigator.language || 'en-US';

let voices = (await spoken.voices()).filter( v => !v.lang.indexOf('en') );
         
// Use the First Voice

spoken.say( 'Welcome to flavor town.', voices[0] );


              
// Speak with each English voice to sample your favorites

(await spoken.voices())

    .filter( voice => voice.lang.indexOf('en') == 0 )

    .map( voice => voice.name )

    .forEach( voice =>

        spoken.say( 'Welcome to flavor town.', voice ).then(

            speech => console.log(speech.voice.name)

        )

    );

Step 5Listening
Writing a document with your voice or chatting with your friend hands-free is possible when you enable continuous listening mode. First you should check to see if the browser has access to the Speech-to-Text capability.


const available = spoken.listen.available();

if (available) console.log('Hurray voice transcription is available!');


// Async/Await

console.log(await spoken.listen());



// or Promise style

spoken.listen()

.then( transcript => console.log(transcript))

.catch(     error => console.warn(error.message))


// We can only have one concurrent listener.

// So you can catch if there is an error.

spoken.listen()


    .then( transcript => console.log(transcript))

    .catch(     error => console.warn(error.message))

Capture live "real-time" transcription as you speak.


              spoken.listen.on.partial( ts => console.log(ts) );


spoken.listen()
    .then(     ts => console.log("Partial: " + ts))

    .catch( error => console.warn(error.message))

Additional voice transcription events.


              spoken.listen.on.start( voice => { console.log('Started Listening') } );

spoken.listen.on.end(   voice => { console.log('Ended Listening')   } );

spoken.listen.on.error( voice => { console.log('Error Listening')   } );

Stop listening.


              spoken.listen.stop();

Continuous Listening


              
spoken.listen.on.end(continueCapture);

spoken.listen.on.error(continueCapture);


              
startCapture();


              
function startCapture() { 

    spoken.listen({ continuous : true }).then( transcript =>

        console.log("Captured transcript: " + transcript)

    ).catch( e => true );

}


              
async function continueCapture() {

    await spoken.delay(300);

    if (spoken.recognition.continuous) startCapture();

}


              
function stopCapture() {

    spoken.recognition.continuous = false;

    spoken.listen.stop();

}