API Change Log

To further speed up Core API, it now supports the following two parameters -
fix_orientation: you can disable this parameter to gain an average of 700ms performance improvement however you should ensure that the image submitted is in an upright position.
disable_barcode: If the document uploaded contains a barcode, Core API will attempt to read the barcode which takes average of 2500ms, if you don't need any information from barcode you may set this option to true.

DocuPass now allows cropping of captured document similar to Core API through the following parameter when creating DocuPass sessions: crop_document.

DocuPass now allows you to directly send your users with an SMS containing the verification link using sms_verification_link parameter. You no longer need to display QR code or send the DocuPass URL to your users with this option enabled, and you can also verify your user's mobile number as part of the onboarding process, since they will not be able to begin verification unless they have provided a correct number to receive the link.

We have released AML compliance solutions which allow our customers to perform comprehensive KYC against their users.

The new AML solution can be accessed via 3 methods:

• AML API - Directly search the AML database using name or document number
• Core API - Automatically check extracted personal information from user document against AML database
• DocuPass API - Automatically check user submitted ID against AML database

We have integrated phone number verification into DocuPass, once phoneverification parameter is enabled, users will be asked to enter a phone number as the first step of verification. DocuPass will then SMS or call this number giving user a verification code which the user will have to enter to pass phone verification.

If you already have your user's phone number, you can also pass it to verify_phone parameter, your user will not be able to change this phone number and must pass the validation based on the phone number you provide.

DocuPass will returned verified phone number, and number type "mobile" or "landline" in the callback JSON.

It is now possible to replace the original DocuPass UI design with your own design using HTML and CSS, check DocuPass API reference for the exact parameter and template download.

We have added two additional checks performed by authentication module v2:

• Security Feature Check - Check if hologram, watermark and laser engravings exist on the document. (Currently only support some US states, but more will be added)
• Recapture Check - Check if the document is taken from a screen (Still under tuning for better accuracy)

To make it easier to parse breakdown results programmatically, an integer code will now come with failed checks, a severity score ranging from low, medium, high has also been added. Read Core API Reference for more details.

We have updated the developer documentations, with this update, we have also released client library for PHP, .NET, Python and NodeJS.

We have made some adjustments to DocuPass callbacks, by setting return_type to 1 you can now receive user document and face images through an URL instead of base64 encoded content. This change allows customers to receive the images without configuring their server to accept a large post request.

To allow further customization of DocuPass modules, it is now possible to set customized welcome message using welcomemessage parameter, providing a customized welcome text will replace the default generic welcome message.

We've added support for our users to customize the speed and accuracy of Core API, by specifying the accuracy parameter you can now trade document recognition accuracy for a faster response speed. This parameter comes with 3 levels: 0 for lowest accuracy but fastest, to 2 with best accuracy but slower performance.

You can now set the cutoff ratio for biometric verifications in both Core API and DocuPass using biometric_threshold.

Traditionally, when submitting a dual-side document to Core API or DocuPass, our system will automatically check if the information(name, document number, DOB etc.) matches between the two sides of the document. This is causing trouble for documents in specific regions, which has different document number between the front and the back of the document. It is now possible disable such check and return a result by setting dualsidecheck to value false.

For DocuPass callback, the API will no longer return base64 encoded video content when video biometric verification is enabled. The video is generally a huge file causing a large POST body size, hence generating error on some of our customers' servers. Instead, we will now extract a face photo from the video and include it in the faceimage list.

It is now also possible to completely disable images being returned by DocuPass callback by specifying the return_documentimage and return_faceimage boolean parameters.

It is now possible to output a cropped image of the face region by specifying true using the outputface parameter.

Web Portal

We have upgrade the your dashboard to a brand new web portal. The new portal allows you to test drive our new products - Vault and DocuPass, along with feature improvements such as subscription plan management and two-factor authentication.

DocuPass

We have released a new identity verification solution called DocuPass, using DocuPass you will be able to onboard your users using our web interface without developing your own document collection process. There are 3 ways to utilize DocuPass, you can either embed the identity verification page as an iframe on your website, or redirect your users to the identity verification page, or you may also choose to do mobile verification where your users will scan a QR code to open mobile DocuPass app. Once identity check is completed, the identity data will be sent back to your server and you can decide how to use them. To learn more about DocuPass, check out the DocuPass Documentation or the DocuPass page under your web portal.

Vault

We have deprecated the previous logging mechanism where you can see API responses under your account dashboard if logging is turned on, instead, we introduce a document database called Vault. Each entry under vault represents an identity, which contains personal information and document and face images. Saving data to Vault is optional, but doing so allows you to quickly search users onboarded, using keywords or even a face image of the person. To learn more about DocuPass, check out the Vault Documentation or the DocuPass page under your web portal.

Video Biometric Verification

It is now possible to upload a video to our Core API for biometric verification, the API will check liveness of the person inside video and produce a similarity score between the face in the video and the document. To further secure your onboarding process, the user must pronounce certain passcode inside the video, the API will analyze the soundtrack and check whether the passcode match. Check out the Video Biometric Documentation.

Authentication v2

We have released a new authentication module uses different matrices to validate documents, to use the newer module you need to supply parameter authenticate_module=2. The v2 module is more verbose and gives insight on what is wrong with the document, along with a score similar to v1.

Dual-side Scan

It is now possible to submit both the front and the back of a document in one API call using file_back, file_back_base64 or url_back. The API will automatically aggregate data on both side and return a single response containing both fields recognized from the front and the back of the document image. Note that dual-side scan costs 2 quota unit.

Minor Changes

- New parameter ocr_scaledown has been introduced to let you fine-tune OCR accuracy and speed.
- log has been deprecated and replaced by vault_save and few other relevant vault related parameters.
- New parameter checkblocklist to check newly uploaded document against a blocked document in vault.

API will now include a field matchrate with every successful request, this value represents the percentage of document data fields successfully parsed by the API.

For example, a driver licence normally has 4 fields: firstname, lastname, address, and date of birth. A document submitted is too blurry, and our API could only recognize firstname, lastname and address, but not the date of birth, the matchrate would be 0.75.

This new field can be used to determine the image quality of the document, if the document is too blurry, cropped the score would be low. A healthy score would be > 0.8 to 0.9. Note that it is normal not to score a 1 for matchrate, there are some optional fields which isn't available on everyone's document.

This update aims at improving compatibility and ease of integration when it comes to uploading image to our service. Traditionally, developers are required to submit a multipart/form-data POST request to the api endpoint, which wasn't a straight-forward process for many programming framework and libraries. Today, without deprecating the aforementioned method, we are adding two extra request fields, file_base64 and face_base64, allowing developers to upload image using common key-value pairs POST type application/x-www-form-urlencoded.

To submit document or face image under the new fields, simply read and encode the entire image file in base64 format. For example in PHP:

 base64_encode(file_get_contents($pathToDocumentImage)),
    'apikey' => 'YOUR API KEY'
];
$ch = curl_init();
curl_setopt($ch, CURLOPT_URL, 'https://api.idanalyzer.com');
curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
curl_setopt($ch, CURLOPT_POSTFIELDS, http_build_query($post));
$response = curl_exec($ch);
var_export($response);

It is also possible to submit the base64 data that comes from HTML img src tag, similar to:

data:image/jpeg;base64,/9j/4AAQSkZJRgABAQE.................

We hoped that the addition of these two parameters will simplify the process connecting your application to our API.

We have updated the API endpoint of ID Analyzer to https://api.idanalyzer.com from https://www.idanalyzer.com/api, the new endpoint has reduced latency, improved security and performance through global CDN.

The deprecated endpoint will continue to function until 31 December 2020, all customers should have sufficient time to make the switch.

We are continuously updating our algorithms to achieve faster response time on ID Analyzer. Today, we have added a new response parameter executionTime, which output a floating number in seconds on the time taken to process your document. Customers can make use of this value to cross reference the settings and benchmark our API.

The key to fast response time is always scaling down the document image before submitting to our API. Scaling down the input image while maintaining text readability should keep the executionTime well under 1.5 seconds, and of course, it will take more time to authenticate, crop, or perform biometric verification.