# WebSDK

# Introduction

The easiest way to start with our solution is to use our WebSDK. We put lots of efforts to improve conversion of your users and we thought about little details that make a difference. You will love WebSDK because:

  • It is highly customizable (choose any steps you want, change styles or copy)
  • Feature a face liveness module that improves conversion and minimizes fraud
  • Adaptive design and mobile friendly
  • Gives hints to users about how to better pass the flow
  • Ability to continue on mobile
  • Automatically refreshes when status is changed

WebSDK Screen

# Integration

We used to have a more complicated way of initializing. If you just started the integration and cannot migrate to the new way, please, contact us - the old way is still supported.

# Steps overview

Here are 3 simple steps that you should do:

  • Step 1: [Dashboard] Set up your applicant flow or pick a predefined one (e.g. basic-kyc)
  • Step 2: [Backend] Generate an Access token to pass it later on to the WebSDK
  • Step 3: [Frontend] Install an NPM package or use CDN version of the script and launch it with the flow you picked on the first step

# Generating an access token

All API requests must be authenticated as described in this section.

To initialize a WebSDK you must generate a temporary access token, which is also passed to the WebSDK initialization. This action must be performed on the server side. You also have to provide a userId query parameter, that is unique and meaningful. It can be an external user ID in your system, or an email address. Do not randomly generate these IDs, unless you do testing.

If you reuse access tokens make sure to set up an access token expiration handler (see the Frontend integration section)

# Frontend integration

# 1a) Install the package - NPM

npm i @sumsub/websdk --save
# OR
yarn add @sumsub/websdk

Import the module

import snsWebSdk from '@sumsub/websdk';

# 1b) Install the package - Standalone

# 2) Create a container for the WebSDK on your page

# 3) Initialize a WebSDK

 * @param apiUrl - 'https://test-api.sumsub.com' (sandbox)
                    or 'https://api.sumsub.com' (production)
 * @param flowName - the flow name chosen at Step 1 (e.g. 'basic-kyc')
 * @param accessToken - access token that you generated on the backend in Step 2
 * @param applicantEmail - applicant email (not required)
 * @param applicantPhone - applicant phone, if available (not required)
 * @param customI18nMessages - customized locale messages for current session (not required)
function launchWebSdk(apiUrl, flowName, accessToken, applicantEmail, applicantPhone, customI18nMessages) {
    let snsWebSdkInstance = snsWebSdk.Builder(apiUrl, flowName)
            (newAccessTokenCallback) => {
                // Access token expired
                // get a new one and pass it to the callback to re-initiate the WebSDK
                let newAccessToken = this.newAccessToken(); // get a new token from your backend
            lang: 'en', //languge of WebSDK texts and comments (ISO 639-1 format)
            email: applicantEmail,
            phone: applicantPhone,
            i18n: customI18nMessages,
            onMessage: (type, payload) => {
                // see below what kind of messages the WebSDK generates
                console.log('WebSDK onMessage', type, payload)
            uiConf: {
                customCss: "https://url.com/styles.css"
                // URL to css file in case you need change it dynamically from the code
                // the similar setting at Applicant flow will rewrite customCss
                // you may also use to pass string with plain styles `customCssStr:`
            onError: (error) => {
                console.error('WebSDK onError', error)

    // you are ready to go:
    // just launch the WebSDK by providing the container element for it

Please make sure that an accessToken was provided to SDK initialization function (it has a format like _act-b8ebfb63-5f24-4b89-9c08-5bbabeec986e and can be obtained with API request) and not other kind of token.

# Standalone working example

Below is a fully working example. Just substitute $FLOW_NAME and $ACCESS_TOKEN with the appropriate values (see steps above), save it to an HTML file and open it in your browser.

# WebSDK messages

When providing the onMessage handler you will receive messages from the WebSDK (all of them are prefixed with idCheck.):

You can see all of those messages in your Browser dev console if you are launching the WebSDK from our dashboard.

messageType Payload Description
// id of the iframe in the current context 
WebSDK resources have been loaded
The first screen is rendered
"idDocSetType": "$idDocSetType",
"types": ["$idDocType1", "$idDocType2"]
A screen with that corresponds to '$idDocSetType' was shown
{"step": "$idDocSetType"}
Step '$idDocSetType' has been completed
{"applicantId": "$id"}
Applicant with id $id has been loaded
Documents were submitted for verification
"reprocessing": false,
"levelName": "$levelName",
"createDate": "$createDate",
"expireDate": "$expireDate",
"reviewStatus": "$reviewStatus",
"autoChecked": false
Applicant status has been changed
Documents were re-submitted for verification
Applicant action was submitted
"applicantActionId": "$id"
Applicant action $id was completed
"answer": "GREEN"
Result of standalone module has been presented to the user. Valid only for "module" flow types. Happens either when the user has just completed the check or when user reopened the module for the previously completed check. Possible answers: GREEN - check was successful; YELLOW - user is allowed to proceed, final check result will be determined later; RED - user has been decisively rejected and is not allowed to proceed. Note, that if the rejection is not final, this even will not fire.
"height": 159
WebSDK frame has been resized
Video call was started by the user
Video call was answered by operator
Video call was completed by operator
"code": "$error",
"msg": "Error message"
Uploaded document was rejected. Available values of errors
"code": "$warning",
"msg": "Warning message"
There are warnings about the uploaded document. Available values of warnings

# Receiving verification results

On the test environment, checks will not be started automatically and therefore the webhooks will not be sent automatically either. Either ask us to do so manually, or trigger a webhook yourself by calling an API endpoint described here. Once integration is done, and you have moved to the production environment, the webhooks will be sent automatically once the verification procedure has been completed.

Your backend receives results via a web-hook mechanism. You can get the results via email, Slack or Telegram. But for production use we recommend setting up a web-hook endpoint that we would call when necessary:

  • We need a webhook URL from you to send your results to. You can set up webhooks in developer's space at test and production environments.
  • We may send several types of webhooks, but the one that is of interest to you is the one described in the "Getting verification results" section. We encourage you to carefully get accustomed to this section, since there are quite a few options for what the results may look like.
  • If you want to manually check the applicant status you can always do it by this method. However, we don't recommend using polling as base method for getting verification results and rather strongly encourage the webhook mechanism. If for some reason you missed the webhooks, don't worry: we make a note of everything that was attempted to be sent, and can resend failed webhooks at any point in time. If webhook request fails we attempt to resend it four times: after 5 minutes, 1 hour, 5 and 18 hours until request succeeds. Check createdAt field of webhook payload to make sure that you're getting relevant applicant status.
  • Also you can monitor webhook statuses and resend them manually on the API Health page of dashboard on test and production.

Please note that we may send several final webhooks, so be prepared to change the user's status on your side too. A typical example is when a user got approved, but later on he was detected as a fraudster, and therefore we block him with a second "RED" webhook.

# Getting the production key

# How can we get the production key?

In order to ensure the best experience for you and your users, before giving you access to the production environment, we would like to test the flow and pass the verification as your end-user. To do so, we would kindly ask you to provide us with a link to your server.

Ensure that:

  • Integration on the test environment is completed, works and meets your requirements
  • There are no dev console errors or warnings
  • You saved our applicantIds to your database and matched your users
  • You received and successfully processed test webhooks from our side
  • You know how to correlate results received in the webhook payload with your users

Please note that we may send several final webhooks, so be prepared to change the user's status on your side too. A typical example is when a user got approved, but later on he was detected as a fraudster, and therefore we block him with a second "RED" webhook.

# What information should be provided to obtain a production key?

Please provide us with:

  • An email where we can send the production credentials
  • Expected amount of incoming users per hour/day/month

Please note that settings won't be transferred from test environment to production automatically. So don't forget to

  • Set up webhooks at the Dev Space tab of production dashboard
  • Transfer levels and flow from test to production using export -> import functionality
  • Tune regulation rules at Global settings tab of the dashboard
  • Create and use production App Token and Secret Key for API authorization
  • Change baseUrl of your RestAPI requests and SDK initialization parameters from test-api.sumsub.com to api.sumsub.com

# Customization

All the customization is done from our dashboard. You can customize texts, styles (via custom CSS), navigation, etc. To change all SDK texts by locales you should use SDK Translations tab on test and production. See hints and guides in the dashboard for more details.

# Twilio integration

If you want to allow your users to send verification links via SMS to their mobile devices or use you own Twilio credentials for Phone verification step, you have to set it up at the dashboard on production and test and make sure that target regions are enabled in your twilio account here.


# How can we get the applicantId?

Two ways:

  • When user completes a first page an applicant is created. At this moment we send webhook to your endpoint with type applicantCreated. Its payload will contain the applicantId
  • GET /resources/applicants/-;externalUserId=${externalUserId}/one

This endpoint will throw a 404 if no applicant exists with such a user id.

Last Updated: 5/14/2021, 5:15:24 PM