This is a barebones proof-of-concept WebRTC audio chat app built using PeerJS. It uses a simple Node backend that keeps track of peer IDs for each call.
This app uses the new WebRTC APIs to connect directly to other users' browsers. Here's how it all works.
When you land on the call page, your browser will prompt you to allow the page to access your microphone. This is done by calling navigator.getUserMedia().
After allowing microphone access, you have access to a LocalMediaStream. There lots of other things you can do with this stream using the Web Audio API, but we are't doing anything fancy here.
Here's what we're doing:
// handle browser prefixes
navigator.getUserMedia = (navigator.getUserMedia || navigator.webkitGetUserMedia || navigator.mozGetUserMedia || navigator.msGetUserMedia);
// Get access to microphone
navigator.getUserMedia (
// Only request audio
{video: false, audio: true},
// Success callback
function success(localAudioStream) {
// Do something with audio stream
},
// Failure callback
function error(err) {
// handle error
}
);PeerJS takes care of the hairier parts of using WebRTC for us (STUN, TURN, signaling). To connect, you need to include the PeerJS javascript, then make a new instance of Peer:
var me = new Peer({key: API_KEY});
me.on('open', function() {
console.log('My PeerJS ID is:', me.id);
});The open event will fire once you're connected.
For this demo, we're sending this ID to the server, where it's associated with this call ID. When someone else opens the call we then pass them everyone else's PeerJS IDs.
When other people open the call page, they'll have your PeerJS ID (and maybe other people's). These IDs are then used to connect to each other.
Each Peer instance has a .call() method that takes a peer's ID, your LocalMediaStream as arguments and returns a PeerJS MediaConnection object.
var outgoing = me.call(peerId, myAudioStream);This MediaConnection object will fire a stream event when the other person answers your call with their own audio stream.
outgoing.on('stream', function(stream) {
// Do something with this audio stream
});When receiving a .call(), your Peer instance will fire a call event, which gets passes an instance of a PeerJS MediaConnection. You then listen for the stream event on this object to get incoming audio stream:
me.on('call', function(incoming) {
incoming.on('stream', function(stream) {
// Do something with this audio stream
});
});Once all this happens, both parties should have an audio stream from the other person.
There are two ways to play an audio stream: the Web Audio API, and the HTML5 <audio> element. Firefox supports both, but Chrome currently doesn't support playing a WebRTC stream using the Web Audio API, so we're using the <audio> element here:
function playStream(stream) {
var audio = $('<audio autoplay />').appendTo('body');
audio[0].src = (URL || webkitURL || mozURL).createObjectURL(stream);
}We use the new URL.createObjectURL() method to get a URL that the <audio> element can stream.
Also note the autoplay attribute. You don't have to have this, but if you don't use it you'll need to manually call the .play() on the <audio> element.
That's it. You can read more about using PeerJS in their documentation.
To run this demo on your computer, first you'll need to get a PeerJS API key (it's free).
Once you have an API key, copy (or rename) config.js.example to config.js, then open it and add your API key there.
Now just install dependencies and run the server:
$ npm install
$ node .
Audio conference using WebRTC and PeerJS.
04e2184d07dec3707760274c533f85f05063cfdf