Skip to content
Go to file


Build Status

The Amazon Connect Streams API (Streams) gives you the power to integrate your existing web applications with Amazon Connect. Streams lets you embed the Contact Control Panel (CCP) UI components into your page, and/or handle agent and contact state events directly giving you the power to control agent and contact state through an object oriented event driven interface. You can use the built in interface or build your own from scratch: Streams gives you the choice. This library must be used in conjunction with amazon-connect-chatjs in order to utilize Amazon Connect's Chat functionality.

Learn More

To learn more about Amazon Connect and its capabilities, please check out the Amazon Connect User Guide.


Run the Makefile to generate amazon-connect-${version}.js, then copy this file into your application or host it in an Amazon S3 bucket behind Amazon Cloudfront.

$ make

Important Announcements

  1. July 2020 -- We recently changed the new, omnichannel, CCP's behavior when it encounters three voice-only agent states: FailedConnectAgent, FailedConnectCustomer, and AfterCallWork.
    • FailedConnectAgent -- Previously, we required the agent to click the "Clear Contact" button to clear this state. When the agent clicked the "Clear Contact" button, the previous behavior took the agent back to the Available state without fail. Now the FailedConnectAgent state will be "auto-cleared", much like FailedConnectCustomer always has been.
    • FailedConnectAgent and FailedConnectCustomer -- We are now using the contact.clear() API to auto-clear these states. As a result, the agent will be returned to their previous visible agent state (e.g. Available). Previously, the agent had always been set to Available as a result of this "auto-clearing" behavior. Note that even custom CCPs will behave differently with this update for FailedConnectAgent and FailedConnectCustomer.
    • AfterCallWork -- As part of the new contact.clear() behavior, clicking "Clear Contact" while in AfterCallWork will return the agent to their previous visible agent state (e.g. Available, etc.). Note that custom CCPs that implement their own After Call Work behavior will not be affected by this change.
      • We are putting contact.complete() on a deprecation path. Therefore, you should start using contact.clear() in its place. If you want to emulate CCP's After Call Work behavior in your customer CCP, then make sure you use contact.clear() when clearing voice contacts.

Getting Started

Upgrading to the OmniChannel CCP (AKA CCPv2)?

If you are migrating to the new CCP, we encourage you to upgrade to the latest version of this repository. You should also upgrade to the latest version of RTC-JS as well, if you are using it. For a complete migration guide to the new CCP, and to fully understand the differences when using Streams with the new CCP, please see this post:


The first step to using Streams is to allowlist the pages you wish to embed. For our customer's security, we require that all domains which embed the CCP for a particular instance are explicitly allowlisted. Each domain entry identifies the protocol scheme, host, and port. Any pages hosted behind the same protocol scheme, host, and port will be allowed to embed the CCP components which are required to use the Streams library.

To allowlist your pages:

  1. Login to your AWS Account, then navigate to the Amazon Connect console.
  2. Click the instance name of the instance for which you would like to allowlist pages to load the settings page for your instance.
  3. Click the "Application integration" link on the left.
  4. Click "+ Add Origin", then enter a domain URL, e.g. "", or "" if your website is hosted on a non-standard port.

A few things to note:

  • Allowlisted domains must be HTTPS.
  • All of the pages that attempt to initialize the Streams library must be hosted on domains that are allowlisted as per the above steps.
  • All open tabs that contain an initialized Streams library or any other CCP tabs opened will be synchronized. This means that state changes made in one open window will be communicated to all open windows.
  • Using multiple browsers at the same time for the same connect instance is not supported, and causes issues with the rtc communication.

Downloading Streams with npm

npm install amazon-connect-streams

Importing Streams with npm and ES6

import "amazon-connect-streams"; This will make the connect variable available in the current context.

Usage with TypeScript

amazon-connect-streams is compatible with TypeScript. You'll need to use version 3.0.1 or higher:

import "amazon-connect-streams";

connect.initCCP({ /* ... */ });

Downloading Streams from Github

The next step to embedding Connect into your application is to download the Streams library from GitHub. You can do that by cloning our public repository here:

$ git clone

Once you download streams, change into the directory and build it:

$ cd amazon-connect-streams
$ make

This will generate a file called connect-streams-${VERSION}.js, this is the full Connect Streams API which you will want to include in your page. You can serve connect-streams-${VERSION}.js with your web application.

Build your own with NPM

Install latest LTS version of NodeJS

$ npm install -g gulp
$ git clone
$ cd amazon-connect-streams
$ npm install
$ npm run release

Find build artifacts in release directory - This will generate a file called connect-streams.js and the minified version of the same connect-streams-min.js - this is the full Connect Streams API which you will want to include in your page.

To run unit tests specifically:

$ npm run gulp-test

Using the AWS SDK and Streams

Streams has a "baked-in" version of the AWS-SDK in the ./src/aws-client.js file. Make sure that you import Streams before the AWS SDK so that the AWS object bound to the Window is the object from your manually included SDK, and not from Streams.


Initializing the Streams API is the first step to verify that you have everything setup correctly and that you will be able to listen for events.


<!DOCTYPE html>
<meta charset="UTF-8">
    <script type="text/javascript" src="amazon-connect-1.4.js"></script>
  <!-- Add the call to init() as an onload so it will only run once the page is loaded -->
  <body onload="init()">
    <div id="containerDiv" style="width: 400px;height: 800px;"></div>
    <script type="text/javascript">
      var instanceURL = "";
      // initialize the streams api
      function init() {
        // initialize the ccp
        connect.core.initCCP(containerDiv, {
          ccpUrl: instanceURL,            // REQUIRED
          loginPopup: true,               // optional, defaults to `true`
          loginPopupAutoClose: true,      // optional, defaults to `true`
          loginOptions: {                 // optional, if provided opens login in new window
            autoClose: true,              // optional, defaults to `false`
            height: 600,                  // optional, defaults to 578
            width: 400,                   // optional, defaults to 433
            top: 0,                       // optional, defaults to 0
            left: 0                       // optional, defaults to 0
          region: "eu-central-1",         // REQUIRED for `CHAT`, optional otherwise
          softphone: {                    // optional
            allowFramedSoftphone: true,   // optional
            disableRingtone: false,       // optional
            ringtoneUrl: "./ringtone.mp3" // optional

Integrates with Connect by loading the pre-built CCP located at ccpUrl into an iframe and placing it into the containerDiv provided. API requests are funneled through this CCP and agent and contact updates are published through it and made available to your JS client code.

  • ccpUrl: The URL of the CCP. This is the page you would normally navigate to in order to use the CCP in a standalone page, it is different for each instance.
  • region: Amazon connect instance region. ex: us-west-2. only required for chat channel.
  • loginPopup: Optional, defaults to true. Set to false to disable the login popup
    which is shown when the user's authentication expires.
  • loginOptions: Optional, only valid when loginPopup is set to true. Provide an object with the following properties to open loginpopup in a new window instead of a new tab.
    • autoClose: Optional, defaults to false. Set to true to automatically close the login popup after the user logs in.
    • height: This allows you to define the height of the login pop-up window.
    • width: This allows you to define the width of the login pop-up window.
    • top: This allows you to define the top of the login pop-up window.
    • left: This allows you to define the left of the login pop-up window.
  • loginPopupAutoClose: Optional, defaults to false. Set to true in conjunction with the loginPopup parameter to automatically close the login Popup window once the authentication step has completed. If the login page opened in a new tab, this parameter will also auto-close that tab. This can also be set in loginOptions if those options are used.
  • loginUrl: Optional. Allows custom URL to be used to initiate the ccp, as in the case of SAML authentication.
  • softphone: This object is optional and allows you to specify some settings surrounding the softphone feature of Connect.
    • allowFramedSoftphone: Normally, the softphone microphone and speaker components are not allowed to be hosted in an iframe. This is because the softphone must be hosted in a single window or tab. The window hosting the softphone session must not be closed during the course of a softphone call or the call will be disconnected. If allowFramedSoftphone is true, the softphone components will be allowed to be hosted in this window or tab.
    • disableRingtone: This option allows you to completely disable the built-in ringtone audio that is played when a call is incoming.
    • ringtoneUrl: If the ringtone is not disabled, this allows for overriding the ringtone with any browser-supported audio file accessible by the user.

A few things to note:

  • You have the option to show or hide the pre-built UI by showing or hiding the containerDiv into which you place the iframe, or applying a CSS rule like this:
#containerDiv iframe {
  display: none;
  • The pre-built CCP UI is portrait oriented and capable of contracting horizontally to fit smaller widths. It can expand from a width of 200px to a maximum of 320px based on the size of its container div. If the CCP is placed into a container where it would be sized under 221px in width, the CCP switches to a different layout style with smaller buttons and fonts.
  • In its normal larger style, the CCP is 465px tall. In its smaller form, the CCP is reduced to 400px tall.
  • CSS styles you add to your site will NOT be applied to the CCP because it is rendered in an iframe.
  • If you are trying to use chat specific functionalities, please also include ChatJS in your code. We omit ChatJS from the Makefile so that streams can be used without ChatJS. Streams only needs ChatJS when it is being used for chat. Note that when including ChatJS, it must be imported after StreamsJS, or there will be AWS SDK issues (ChatJS relies on the ConnectParticipant Service, which is not in the Streams AWS SDK).
  • If you'd like access to the WebRTC session to further customize the softphone experience you can use connect-rtc-js. Please refer to the connect-rtc-js readme for detailed instructions on integrating connect-rtc-js with Streams.

Where to go from here

Check out the full documentation here to read more about subscribing to events and enacting state changes programmatically.

You can’t perform that action at this time.