View on GitHub


a bunch of vanilla JS practice projects based on JavaScript 30 by WesBos.


demo gif

Back to home pagev

User stories


Connect the mic and the browser

As previous tutorial, we need a server for connection this time. As I have no idea how server works, I followed the instruction of including the package.json file and type npm install to start it. I received a message for fixing errors in my shell(Git Bash) after typing npm install, I followed the error instruction, typed npm install again and it worked.

Unlike the tutorial, my voice note works just fine in Chrome.

window.SpeechRecognition = window.SpeechRecognition || window.webkitSpeechRecognition; //only work in firefox

//create instance of SpeechRecognition
const recognition = new SpeechRecognition();
//populate result simultaneously(as user talk)
recognition.interimResults = true;

“Write” with voice

Every p will be store inside a <div class="div_notes" contenteditable></div> using appendChild. We will put the result we get into the p and generate a new one if user stop talking, make sure to restart the recognition every time it ends.

We can also detected what user said and take action, in this build we reset our notes.

let p = document.createElement("p"); //create new p every time
const words = document.querySelector(".div_notes");

recognition.addEventListener("result", e => {
 const transcript = Array.from(e.results)
  //map to the transcripts and join all of those together
  .map(result => result[0])
  .map(result => result.transcript)

 p.textContent = transcript; //place what user said in p
 //create a new p whenever user stop
 if (e.results[0].isFinal) {
  p = document.createElement("p");

 if (transcript.includes("reset notes")) {
  words.innerHTML = "";

recognition.addEventListener("end", recognition.start); // start again when it ends