HTML5 Speech Input with Voice Recognition

HTML5 support speech input (x-webkit-speech) and this has been implemented in the newer versions of Google Chrome. Dictation is an online speech recognition app that uses Google’s Chrome speech engine to help you convert your spoken words into text.

This piece of jQuery / JavaScript code powers the Dictation app.

// Written by Amit Agarwal on 10/08/2012
// See for live demo

$(document).ready(function () {

// Check if the user's web browser supports HTML5 Speech Input API
  if(document.createElement('input').webkitSpeech == undefined) {
    $(".answer").append("We are sorry but Dictation requires Google Chrome.");
  else {

// Get the default locale of the user's browser (e.g. en-US, or de)
    var language = window.navigator.userLanguage || window.navigator.language;
    $("#speech").attr("lang", language).focus();

// Make the text region editable to easily fix transcription errors
    $(".answer").click(function () {
       $('.answer').attr('contentEditable', 'true');

// This is called when Chrome successfully transcribes the spoken word
  $("#speech").bind("webkitspeechchange", function (e) {
     var val = $(this).val();

// Did the user say Delete? Then clear the canvas.
     if(val == "delete everything") {
// For "new line" commands, add double line breaks.
     if(val == "new line") 
        val = "<br /><br />";
     else {

// Capitalize the first letter of the sentence.
	val = val.substr(0, 1).toUpperCase() + val.substr(1);

// If the last letter is a alphanumeric character, add a period (full stop)
           val = val + ".";

// Append the transcribed text but set the focus to the hidden speech input.
// This enables keyboard shortcut Ctrl+Shift+Period (.) for speech mode.
     $(".answer").append(val + " ").fadeIn();