Open VoIP Alliance Webphone Lib

Makes calling easier by providing a layer of abstraction around SIP.js. To figure out why we made this, read our blog post.

Documentation

Check out the documentation here.

Cool stuff

• Allows you to switch audio devices mid-call.
• Automatically recovers calls on connectivity loss.
• Offers an easy-to-use modern javascript api.

Join us!

We would love more input for this project. Create an issue, create a pull request for an issue, or if you're not really sure, ask us. We're often hanging around on discourse. We would also love to hear your thoughts and feedback on our project and answer any questions you might have!

Getting started

$ git clone [email protected]:open-voip-alliance/WebphoneLib.git
$ cd WebphoneLib
$ touch demo/config.mjs

Add the following to demo/config.mjs

export const authorizationUserId = <your-voip-account-id>;
export const password = '<your-voip-password>';
export const realm = '<realm>';
export const websocketUrl = '<websocketUrl>';

Run the demo-server:

$ npm i && npm run demo

And then play around at http://localhost:1235/demo/.

Examples

Connecting and registering

import { Client } from 'webphone-lib';

const account = {
  user: 'accountId',
  password: 'password',
  uri: 'sip:accountId@<realm>',
  name: 'test'
};

const transport = {
  wsServers: '<websocket-url>', // or replace with your
  iceServers: [] // depending on if your provider needs STUN/TURN.
};

const media = {
  input: {
    id: undefined, // default audio device
    audioProcessing: true,
    volume: 1.0,
    muted: false
  },
  output: {
    id: undefined, // default audio device
    volume: 1.0,
    muted: false
  }
};

const client = new Client({ account, transport, media });

await client.register();

Incoming call

// incoming call below
client.on('invite', (session) => {
  try {
    ringer();

    let { accepted, rejectCause } = await session.accepted(); // wait until the call is picked up
    if (!accepted) {
      return;
    }

    showCallScreen();

    await session.terminated();
  } catch (e) {
    showErrorMessage(e)
  } finally {
    closeCallScreen();
  }
});

Outgoing call

const session = client.invite('sip:518@<realm>');

try {
  showOutgoingCallInProgress();

  let { accepted, rejectCause } = await session.accepted(); // wait until the call is picked up
  if (!accepted) {
    showRejectedScreen();
    return;
  }

  showCallScreen();

  await session.terminated();
} catch (e) {
} finally {
  closeCallScreen();
}

Attended transfer of a call

if (await sessionA.accepted()) {
  await sessionA.hold();

  const sessionB = client.invite('sip:519@<realm>');
  if (await sessionB.accepted()) {
    // immediately transfer after the other party picked up :p
    await client.attendedTransfer(sessionA, sessionB);

    await sessionB.terminated();
  }
}

Audio device selection

Set a primary input & output device:

const client = new Client({
  account,
  transport,
  media: {
    input: {
      id: undefined, // default input device
      audioProcessing: true,
      volume: 1.0,
      muted: false
    },
    output: {
      id: undefined, // default output device
      volume: 1.0,
      muted: false
    }
  }
});

Change the primary I/O devices:

client.defaultMedia.output.id = '230988012091820398213';

Change the media of a session:

const session = await client.invite('123');
session.media.input.volume = 50;
session.media.input.audioProcessing = false;
session.media.input.muted = true;
session.media.output.muted = false;
session.media.setInput({
  id: '120398120398123',
  audioProcessing: true,
  volume: 0.5,
  muted: true
});

Commands

Generate documentation

Typedoc is used to generate the documentation from the jsdoc comments in the source code. See this link for more information on which jsdoc tags are supported.

Run puppeteer tests

Using docker

Add a .env file with the following:

USER_A = <user-a>
USER_B = <user-b>
PASSWORD_A = <password-user-a>
PASSWORD_B = <password-user-b>
NUMBER_A = <number-user-a>
NUMBER_B = <number-user-b>
WEBSOCKET_URL = <your-websocket-url>
REALM = <realm>

Then call docker-compose up to run the tests.

Note: Don't forget to call npm ci in the puppeteer folder. :)

Without docker

If you don't want to use docker, you will need to run the demo with the npm run demo command (and keep it running) and run the tests with npm run test:e2e. For this you will need the .env file with your settings.