Release notes

Find out what's new in Payment Orchestration:

Dec 2020: 3DS 2 integration

Jan 2021: Soft declines in 3DS 2

Feb 2021: Lightbox in payment page and Mobile SDKs (beta)

March 2021: Domains and Accounts API for Merchants

July 2021: Transaction Export API

Open Payment Gateway - OPG

This documentation guides you through our Payment Orchestration Open Payment Gateway (OPG) features for frontend checkout and backend use cases. It provides important information about integration scenarios, testing possibilities, and references.

For a Glossary of Terms, go to the end of this document. 

First Steps

Every potential checkout webpage includes information about the purchase, payment amount, customer details, plus a few additional details. This information is always issued from your server-side system via HTTP POST Request. This is what we call a LIST Request.

To test these features, you need an account for our website. If you do not have an account, please email welcome.germany@payoneer.com and we will create one for you. You can start experimenting with OPG Sandbox as soon as you get your Payment Orchestration account details.

Try your first LIST Request by following these 4 steps:

• In the portal, go to the Dashboard and open the Provider Contracts section.
• Click on "Add a new contract". This can alternatively be done by clicking on "Add Contract" while in the Dashboard.
• Choose Test Adapter "TESTPSP". Make sure you add a valid URL eg (http://www.dummyvalidurl.com)
• Click on "Select Methods and Countries" under the section "Methods and Countries".
  • In "Step 1: Countries", select Germany and France.
  • In "Step 2: Methods", add payment methods Visa, Mastercard, SEPA Direct Debit and PayPal.
  • In "Step 3: Combinations", select PayPal to be available in France, but not in Germany.
• Access the Routing App from the Merchant Portal, and activate the routes per country by choosing a routing strategy.

You must generate a Token to authenticate your system against the Payment Gateway API.

To generate a Sandbox token:

• Go to the portal dashboard, open "Tokens" and generate a Sandbox Payment Token. It will be used later to authenticate your system against the Payment Gateway.
• Copy and save the value immediately. For security reasons, it will not be shown again and cannot be recovered. You can create a new one if necessary.

We suggest that you use one of the browser plugins listed in Tools for Manual Testing so you can submit JSON POST requests easily by hand.

Use the following values to submit your first LIST Request with the chosen tool:

• URL: https://api.sandbox.oscato.com/api/lists
• HTTP Method: POST
• Basic Authentication User Name: [Your merchant code]
• Basic Authentication Password: [Your sandbox payment token]
• Header Content Type: application/vnd.optile.payment.enterprise-v1-extensible+json
• Header Accept: application/vnd.optile.payment.enterprise-v1-extensible+json
• Request Body Content: Copy and paste the content on the side to the tool that you're using. This content is just an example; make sure that it stays in JSON Format.

Find out more about authentication and version headers in API Access.

For your request you should receive a successful HTTP 200 response with a list of available payment methods for this transaction (attribute networks.applicable).

If you configured your OPG Sandbox as described above, Carte Bleue will not show in this list.

5. Visualize the LIST

Follow the steps below to visualize the content of a generated LIST with a default style, using our Orchestration Platform's Hosted Payment Page:

• Examine the LIST Response and take the content of the links.self attribute. That's the URL of the LIST Resource:

  https://api.sandbox.oscato.com/pci/v1/59722056cb4280f9550078e3lffpph128vffgueil7chn475bs

• Append it as query parameter listUrl to the following URL of the Hosted Payment Page, and open it in a browser:

  https://resources.sandbox.oscato.com/paymentpage/v3/responsive.html?listUrl=https://api.sandbox.oscato.com/pci/v1/59722056cb4280f9550078e3lffpph128vffgueil7chn475bs

Note: You have 30 minutes to make a successful CHARGE. After 30 minutes, you will get an error and you should create a new LIST Session (That's also why the link above will not work).

The payment page should look like the example picture below:

This payment page has all necessary capabilities to validate and process payment account input. For example, try using the following test account and then click on the Pay button:

Card number: 5500000000000004
Expiry Date: any date in the future
Security Code: any 3 digit number
Account Holder: John Doe

Using this input above will trigger instant validation, and clicking on Pay will trigger a frontend CHARGE request and should take you to the success URL defined in the LIST request. Alternatively, you can perform the CHARGE request from your backend which we will simulate together in the next step.

6. CHARGE request

The LIST response as generated in step 4. contains information about the current payment session, including all available payment networks and their attributes as seen in the array networks.applicable. In this example every network has the same attributes, but for our testing we are now only going to look at links.operation from Mastercard:

"operation": "https://api.sandbox.oscato.com/pci/v1/59722056cb4280f9550078e3lffpph128vffgueil7chn475bs/MASTERCARD/charge"

Notice that this operation endpoint starts with a reference to the current LIST long ID and ends in /MASTERCARD/charge, which means that for the Mastercard network the next operation after listing it is a charge on the customer account. This operation URL is the endpoint to which you need to submit the CHARGE request via POST.

In the case of Mastercard, the charge cannot happen without a customer account. So, the CHARGE in this case will carry also a JSON body containing the collected customer account. We can use the same data as done in the step 5. with the hosted payment page, but now mapping them into the appropriate attributes as expected by the payment API.

The CHARGE request will then be done as exemplified in the right pane, using the same authentication and headers as previously done with the LIST request.

A successful CHARGE response should contain a "charged" status code and a redirect URL to the success page as defined in the LIST request.

Make another LIST Request, with an incremented transactionId (it's technically not required but it is good practice to keep them unique) and change the attribute country to FR. Submit again. Now you should see a similar list in the response, but this time with PayPal because we configured it earlier for France only.

Also, you can copy the logo URL (attribute links.logo) of one of the networks and open it in a browser. The logo of that payment method will show.

Take one of the credit cards in the list and open the URL from links.localizedForm in a browser. You will see a snippet of a simple HTML form without any styling. If you like, you can open the HTML source code. This snippet is another building-block provided by OPG to generate the payment page.

To see an out-of-the-box example of how a payment page is rendered, see our Demo. You can also play around with the country parameter that should go into the LIST Request as well as several parameters offered by our AJAX Integration JavaScript library.

From here on we recommend you to understand our Integration Scenarios and choose the one that better fits your business model. We also have more LIST use cases and backend use cases designed to cover specific checkout flows to fit your application.

Tools for Testing

We suggest that you use one of the browser plugins listed below so that you can easily submit JSON POST requests when you test.

• Insomnia
• Postman
• Firefox rester
• Chrome Advanced REST client
• Command line tool: cURL

Get your Merchant Code and Payment Token together, and run this command:

Note that this request does not use JSON as content type for the input but (implicitly) application/x-www-form-urlencoded. If successful, the result will be JSON script and start like this (depending on your configuration):

API Access

API Types and Authentication

In our API Reference you will find a complete and formal documentation of all API calls. Please note the distinction between Client API and Server API. Endpoints in the Client API are designed to be accessed by frontends, eg web browsers, without authentication. For example the CHARGE Request exists in a frontend accessible version. The Server API is for calls from your servers, such as the LIST Request.

Server API calls require HTTPS Basic Authentication with these credentials:

• Login Name: [your merchant code]
• Password: [your payment token]

To generate payment tokens, you access the Merchant Portal (requires access rights). Contact our support team at support.de@payoneer.com for help with permissions.

API Changes

In your contract with us we state that new API response parameters can be added in any release without explicit warning. This will not affect the existing functionality and not require further action from you. Your system only needs to be robust against additional response attributes.

For example, if you use Java with the Jackson parser for JSON, please set the FAIL_ON_UNKNOWN_PROPERTIES to false (see the official Jackson parser documentation and additional details on Stackoverflow).

Of course we will never rename any request or response parameters and never introduce new mandatory request parameters without the contractual notice periods. So with a robust system as described above you are safe.

Accept and Content Type Headers

For every request, you should set two HTTP headers:

• Content Type indicates the content format of your input
• Accept indicates the content format you expect to receive in the reply

For both please use the value: application/vnd.optile.payment.enterprise-v1-extensible+json

This implies communication via JSON. For the level and version part (enterprise-v1-extensible in this case) we may allow custom contractual agreements and hence additional values in the future.

CORS support in Sandbox

Although the Sandbox and Live environment APIs are generally the same, there is one slight technical difference to allow for particular testing and demonstration scenarios:

The Sandbox environment will generally allow Cross-origin resource sharing (CORS), by providing the shown additional HTTP headers in the reply >>

The Live environment only allows CORS  for the very few Client Payment API calls (see above at Authentication and the API Reference). In all other cases CORS will not return the aforementioned headers -- so a LIST Request from a browser (for example) is not possible.

As long as you do not try to sidestep the rules from the Basic Authentication and try to initiate server-to-server calls from a browser, there is nothing to worry about; you will see no difference between Sandbox and Live environment APIs.

DEMO

Integration Scenarios

Simple integration includes the scenarios shown below:

By contrast, advanced integration includes:

Hybrid Scenarios

If you already have integrations with payment providers or are thinking about a parallel setup where the Orchestration Platform-powered payment methods are combined with methods of other sources, there are multiple options to achieve this, depending on your specific prerequisites and needs. Please have a look at our Hybrid & Migration Strategies.

Integration Scenarios Comparison

The Payment Gateway gives merchants a lot of flexibility. It allows multiple ways to integrate the frontend part of the payment processes (compared to backend interactions, which follow a standard path). Each integration scenario has different advantages, regarding effort, conversion, user experience, and security certification.

We generally recommend the AJAX Integration. It provides very good user-experience with minimum implementation effort, and it can be combined with Selective Native integration to minimize PCI requirements on your side. But a different integration scenario will make more sense for some merchant requirements: for integration with mobile devices or other devices that do not rely on web browser rendering eg Smart TVs, we offer special mobile integration scenarios.

Simple Integrations

Hosted

1. You initialize the payment from your backend (LIST Request). Set the parameter style.hostedVersion to v3 to get the most recent version of the hosted payment page (v3). It features a full responsive design and high-resolution payment-method logos. The legacy version (v2) can be accessed by setting style.client to "RESPONSIVE" instead. In the right pane there is an example LIST response demonstrating the location of the URL of a hosted payment page.

2. Based on the information given, OPG generates a web page with all the payment methods that can be used to pay this particular transaction. As a reply, your system gets a URL to this generated page, which it typically displays is a large iFrame.

3. Once the customer finishes their interaction within the iFrame and submits the form, the payment data is sent from within the iFrame directly to the OPG. The response will go back to the iFrame, and as a result the iFrame will redirect its parent frame to the merchant's returnUrl (or cancelUrl in rare error cases) with all available status information attached as query parameters to the URL. The right pane shows an example redirect after success.

Selective Native with AJAX

Selective Native can be implemented using our AJAX Library, or as a native solution. In both cases, the logic behind the implementation is the same; the only difference is that native integration needs the development of your own JavaScript code, while AJAX integration uses a library provided by the Orchestration Platform.

Here's an overview of how our AJAX library communicates with the OPG and especially the Selective Native iFrames. You can use the session "AJAX library Integration" as a reference to our AJAX library.

The Orchestration Platform AJAX Integration Library enables seamless integration of the OPG payment checkout into your own websites, with minimum effort.

Essentially, the JavaScript library takes over a major part of the communication with the Orchestration Platform OPG and creates the actual user interface for checking out. You can read about the general communication flow with OPG in our Getting Started Overview. In the AJAX integration scenario, you only need the initial LIST Request implemented in your server system, and a Notification Listener.

A Selective Native with AJAX integration requires only PCI DSS SAQ A compliance. Download the PCI DSS SAQ A requirements. You must sign the self-assessment questionnaire (SAQ) to confirm that you are compliant. This signed SAQ is also your "Attestation of Compliance (AoC)", which may be requested at any time by your acquiring bank.

Selective Native gives a level of security for credit and debit cards in the payment page that means you need only be PCI A compliant, as opposed to PCI A-EP which requires more effort. Compared to the Hosted scenario, Selective Native integration gives more flexibility over the look and feel of the page -- so you can align it more easily to your corporate design and optimal conversion rates.

Selective Native can be implemented natively on your side, or combined with our own AJAX Library. So if you are using our AJAX library, no extra implementation effort is needed; simply change the integration parameter in the LIST Request to SELECTIVE_NATIVE. For Hosted Integration Selective Native is not necessary, because the whole hosted page already comes from a PCI-certified Orchestration Platform domain.

Selective Native replaces those parts on the payment page with iFrames, where necessary. This means that all credit and debit card forms will be in small iFrames that come from a fully PCI-certified Orchestration Platform domain. The card data entered by a user cannot be read from the surrounding page. That said, all other parts of the page are yours, with the corresponding visual style. Even the small iFrame forms can be styled by providing a CSS file.

This is how the form inside the iFrame would be embedded into the payment page:

Here we've shown you how to handle the Selective Native iFrames at a high level. Use our AJAX Library Integration as a reference if you want to implement this part on your own.

Here is an overview of the client-server and inter-frame communication messages when you use our AJAX library:

You initialize the payment from your backend (LIST Request).

1. The long ID of the generated method LIST is passed back to the customer's browser, with the full list of methods and all resources (images, HTML input forms, etc). At this point, you hand the long ID over to our AJAX Integration JavaScript library, and specify the HTML container element.
2. The JavaScript library now builds the whole payment selection form inside the given HTML container element. This way, it is fully integrated into your website.
3. Your customer can now submit their payment details and be redirected to next page (your success or cancel page, or an external page like PayPal). Your backend will receive a status notification about the payment result.

Note: This approach uses client-side payment page generation. But if you implement this yourself, you could also generate it on the server-side. For more information on how to build this scenario, go to AJAX Integration.

Display Native with AJAX

Here is how communication works in this scenario :

You initialize the payment from your backend (LIST Request).

1. The ID of the generated method LIST is passed back to the customer's browser where you hand it over to our AJAX Integration JavaScript library and specify the HTML container element.
2. The JavaScript library will now automatically get the full list of methods and all resources (images, HTML input forms, etc) and build the whole payment selection form inside the given HTML container element. This way it is fully integrated into your website.
3. Your customer can now submit their payment details and be redirected to the next page -- your success or cancel page, or an external page like PayPal. Your backend will receive a status notification about the payment result in all cases.

For more information on how to build this scenario, go to AJAX Integration.

Advanced Integrations

Selective Native

Selective Native can be implemented as a native solution or by using our AJAX Library. In both cases, the logic behind the implementation is the same; the only difference is that a native integration requires the development of your own JavaScript code, while an AJAX integration uses a library provided by the Orchestration Platform.

For an overview of Selective Native, go to Selective Native with AJAX.

For a detailed explanation of how to implement Selective Native, go to Selective Native Integration.

Display Native

For users without web browser rendering (eg native mobile apps) there is Half-Native Integration a variation of Display Native integration.
Display Native Integration

1. You initialize the payment from your backend (LIST Request) and, based on the information provided, OPG replies with a JSON structure containing all possible payment methods that can be used for this particular transaction (LIST Response). With every network, there is also an HTML snippet representing the corresponding form where the customer needs to fill in their payment account details (eg credit card number etc).
    
2. With this information you can build your own payment page. Your implementation should not be aware of the payment networks in advance. Everything should be fetched and generated at runtime. This way, you can activate additional payment networks by configuration only (ie without changing your implementation).
    
3. The payment account details are then submitted directly to OPG via the CHARGE Request. Since your server does not touch the sensitive payment-account details, you do not have to be PCI compliant.

Redirect Networks such as, for example, Sofortüberweisung do not require payment account details in advance. The corresponding form and the subsequent CHARGE request will be empty in this case. Instead, the redirect URL will point to the external page provided by that network (eg the PayPal login screen). After this step, the customer will be redirected back to your system. For you this is transparent, you do not need to do anything about those networks in your implementation explicitly (see also the details on the CHARGE Request). In this scenario the Status Notification is important because the CHARGE reply goes direct to the client and not to the Merchant's server system.

Read more about how to analyze the LIST Response.

To implement Orchestration Platform's localized forms snippets returned from the original List response, please see Localized Forms section.

Mobile Native

"callback":
{
    "appId": "com.payoneer.checkout.examplecheckout.sdk",
    "notificationUrl": "https://dev.oscato.com/shop/notify.html"
}

Pure Native

Pure Native Integration

Native integration means you build the payment checkout page yourself, based on the information provided by the LIST response, on your server side. It gives you full control and seamless integration into your pages.

In native scenarios, you can send the payment data submitted by your customers, to your own servers first, instead of having it sent directly to OPG. This depends on whether you want to see sensitive payment data and if you are permitted to do so by your PCI DSS certification.

Pure Native integration lets you pass sensitive data through any of your servers that process cardholder data. In this scenario the merchant must have PCI SAQ D certification.

These requirements must be checked, signed, and submitted for revision if necessary.

For the LIST Request and payment-page rendering, this scenario is like the Display Native scenario except that here, the sensitive payment-account details go from the payment checkout page to your own servers first, where they can be stored or analyzed by your custom logic. From there, the CHARGE Request is submitted to OPG.

Note that under credit-card industry regulations, you are only allowed to process credit-card details on your servers if you have the appropriate PCI DSS certification. This covers not only storage of data but also in-memory processing.

Read more about how to analyze the LIST Response.

If you wish to implement Orchestration Platform's localized forms snippets returned from the original List response, go to Localized Forms section.

Pure Native with CSE

The Orchestration Platform will provide you with a public key for encryption in advance. You should store this key because, later on or during the process, the software-client of the user (eg web page in browser, but possibly also mobile clients) will use it.

The payment process is initialized with the LIST Request. You can build the payment page server-side or client-side, just like the other integration scenarios. But, the Orchestration Platform's  AJAX library is not ideal for building the client-side payment page for client-side encryption, since it submits payment data direct to the OPG instead of your custom endpoint. If you want to use it anyway, you can request a development version of the library and then adjust it to your needs.

For this flow to comply with PCI A-EP level compliance, the payment-account data (eg card data, SEPA IBAN etc) entered by the customer must be encrypted by the client so that only the Orchestration Platform can decrypt it. To make that happen, this flow uses the public key referred to above, and the algorithms and data structures described below. Next, you have a choice: you can use the timestamp of encryption and/or the ID of the LIST session for encryption, to assure that the card data is being used for this specific payment session.

In a web context, the Orchestration Platform will provide a JavaScript library providing a function that handles the encryption for you. Your client can then send the encrypted data anywhere to your systems. As your server systems will not be able to decrypt the data without access to the Orchestration Platform's private key, this flow requires only PCI A-EP certification instead of PCI D. Your system will then send the encrypted account-data to the OPG via the CHARGE Request (in the context of the corresponding LIST session).

In this scenario, there is a simultaneous response to your server systems, with all relevant information. The OPG will also send the Status Notification that your server systems would need in other integration scenarios, but it is less significant here because we are already using reliable server-to-server communication. In a web context, your systems then typically respond to the client with an HTTP redirect, depending on the transaction result.

Here is a view of the process:

1. The LIST Request and payment-page rendering are very similar to a Pure Native implementation.
2. But, sensitive account-data are encrypted on the customer-client before being sent to your servers.
3. From there, the CHARGE Request is submitted and the data can only be decrypted by the OPG.
4. Your systems will receive status notifications and redirection URLs as in a standard Pure Native scenario.

Note that under credit-card industry regulations, you must ensure that you have the appropriate PCI DSS certification for processing credit-card details on your servers. This covers not only storage of data, but also in-memory processing. Owing to the nature of the Client-Side Encryption, PCI A-EQ level certification is required.

Read more about how to analyse the LIST Response.

To implement Orchestration Platform's localized forms snippets returned from the original List response, go to Localized Forms.

Half-Native Integration

The Half-Native integration is typically used for devices or applications that do not rely on web-browser rendering, such as native smartphone or smart TV apps. Here, the data from the LIST Response cannot be integrated direct, because the input forms for the payment methods are described in HTML.

As a consequence there could be a translation client-side from the HTML elements into native UI interaction elements. This would preserve the 'implement once' principle but needs a certain budget. Edge cases that need some extra JavaScript logic, such as interactive forms for installment plan selection, are not straight-forward to translate.

The idea of Half-Native therefore is that you, as a merchant developer, prepare the user-interface snippets that correspond to the HTML form snippets provided by the LIST Response in advance. In other words, you prepare all relevant input forms that may occur beforehand natively, eg in terms of the iOS or Android platform.

After your system issued a LIST Request the client will evaluate the response of the client. Then it only shows those natively prepared input forms that were selected by the Payment Gateway for this payment session.

When your customer enters the necessary payment data, the client typically submits it direct to the Payment Gateway via a CHARGE Request -- in the same way as for the Display Native Scenario.  An immediate response and a Status Notification will be sent to your server system.

A corresponding communication sequence looks like this:

Alternatively, the payment session could be initialized as Pure Native and the CHARGE Request issued to your server systems first. But this would affect your required PCI certification.

Another use-case for Half-Native integration is to split the method selection from the actual entering of the account data by the user, ie divide the common payment page into two separate steps.

Here, you could add some simple logic to show the radio buttons and logos on the first part of the payment page, with the second part of the payment page showing the logo and forms of a previously selected payment method with or without the possibility to still switch to a different payment method.

To support this, OPG offers an easy way to store the user selection from the first part in the LIST Session object. This is especially relevant if you want to render the second part as a Hosted Page because it will respect the selected method and show it as preselected (among the other methods).

Every method offered in a LIST Session has the attribute selected and is also a resource with a separate endpoint. Its URL is given as the links.self attribute like this:

If the user selects one method, your system can issue a PUT or PATCH request to that resource to update its selected attribute like this:

Example request

• URL: https://api.sandbox.oscato.com/api/lists/57a06162e4b0a803efa9e1d1l/VISA
• Method: PUT
• Basic Authentication: Not required except for Native Scenarios

The selected status of all other methods in that LIST will become false as a result. So note that other attributes cannot be changed.

JSON Forms

When executing a LIST request (any variation, POST, GET or PUT), the requested URL can contain the parameter view, with options to enable or disable form representation methods.

For example, a LIST request (POST) could look like:

https://api.sandbox.oscato.com/api/lists?view=jsonForms,-htmlForms

With regard to the query parameters, please note:

• If jsonForms is present, the JSON forms are enabled.
• If -htmlForms is present, the HTML forms are not shown. That means: The attributes links.form and links.localizedForm are not present.
• If there are conflicting or unknown arguments like view=jsonForms,-jsonForms or view=abc then the request will fail as invalid.
• If any query parameter is present other than view it will be ignored.

If JSON forms are enabled (view=jsonForms), the LIST response object will contain a form representation in JSON format. This means:

• For each applicable method (network) there is a new attribute structure named localizedInputElements (not inside the links structure, but outside on the same level)
• This structure is a range of objects, and each object represents one input field as follows:
  • name (mandatory): the name of the parameter represented by this input field (as used inside an account object in a consecutive CHARGE request)
  • label (mandatory): the localized human readable label that should be displayed with the input field
  • type (optional): represents the input type / restrictions that can be enforced by the client. These types are allowed:
    • string (default): one line of text without special restrictions. For example: holder name
    • numeric: numbers from 0 to 9, space delimiters and dash ("-") are allowed. For example: card number
    • integer: only numbers from 0 to 9. For example: CVC
    • checkbox: values "true", "false", or undefined (non-existent).
    • select: a list of possible values is given in an additional options attribute (see below). For example: expiration date.
  • options (optional, present for "select" types): a range of objects with attributes:
    • value (mandatory)
    • label (optional. If not present, value should be displayed to the user.)
• If the form definition for a network is not yet present in OPG, an empty object will be returned.

Frontend Checkout

Checkout Flow Overview

The Interactive Checkout Flow Diagram

Especially for developers we created also a systematic view of all combinations of the Orchestration Platform's checkout API flows. Switch between combinations in the selection box at top center. The diagram represents the relationship between a (typical) user journey and the corresponding API requests (with limited details).

Click here to open the Interactive Checkout Flow diagram in a new tab. Loading can take a few seconds, so please be patient...

Regular Checkout

The Orchestration Platform supports two main types of a regular checkout:

• Regular checkout with immediate charge: Your customer arrives at the checkout page via the regular web browser, or via a remote checkout initiated by your Call Center where the list of products to be purchased is displayed. The customer is then directed to the payment page. The payment page is rendered based on how you choose to handle the Orchestration Platform's LIST Request, ie by a simple or advanced integration scenario with the OPG (more information about Integration scenarios). On the payment page your customer enters their payment data and confirms the order. Once the customer confirms the order, the CHARGE takes place and the customer will be redirected to the thank-you page. You can choose to close the sale once you make the order ready for delivery as a deferred charge.
• Regular checkout with Summary Page: Your customer follows the path described above, but the customer is presented with a summary page after they enter their payment details. The summary page gives your customer the opportunity to review their order and check their delivery address and payment details before they confirm the purchase. Presenting the summary page requires a Delayed Payment Submisison (DPS) which involves caching sensitive data, and a few challenges. Your system can temporarily cache payment data via a PRESET Request. Once your customer confirms the order from the summary page, the CHARGE takes place and they will be redirected the thank-you page. You may choose to close the sale once you make the order ready for delivery as a deferred charge.

Express Checkout

• Express Checkout (for first-time customers): On the shopping cart page you already give your customers the option to use a specific Express method such as PayPal Express or Amazon Pay. If they choose one of them then they are directed to that provider or provider interaction elements to make the payment with a user account they already have there. Once the customers confirm on the provider side, their data such as shipping address are shared with your system. Your system would typically present a summary page for the customer to confirm in the fashion of Delayed Payment Submisison (DPS). The detailed Express Checkout documentation you can find in this separate space.
• One-Click Checkout (for returning customers): You can speed up and improve the checkout experience for your return customers by sending these customers direct from the checkout page to the summary page. On the summary page, you show them their registered address and payment details and then the customer just needs to confirm the order. In this process you can perform the Charge by skipping the payment page. Again, you close the sale once you make the order ready for delivery. To offer the One-Click Checkout, you need to give your customers the option to register their account details on their initial purchase. Currently One-Click checkout is currently based on Payment Account Registrations.

Payment Page: LIST Request

A valid LIST Request creates a LIST object. It responds with the applicable payment methods (networks) for this checkout and can also be seen as a kind of payment session of the user in which he or she can do multiple CHARGE attempts.

A typical LIST Request looks like this:

• URL: https://api.sandbox.oscato.com/api/lists
• Method: POST
• Authentication:
  • Login Name: registered company code
  • Password: generated API token

Explanations for some parameters in the request body:

• transactionId is the alphanumeric identifier for this transaction, assigned by you as a merchant.
• country is the "transaction country". It specifies which country-specifc OPG configuration and therefore payment method selection should be used. Typically it is the location of the current shop but it may also be the origin of the customer. Country codes according to ISO 3166-1 alpha-2 are used.
• integration defines which integration scenario the payment session is initialized for, see also below. Valid values are: HOSTED, SELECTIVE_NATIVE, DISPLAY_NATIVE (default) and PURE_NATIVE.
• customer.number is the alphanumeric identifier of your customer, assigned by you as a merchant.
• payment.reference is the short text that customers will see on their account statement.
• style.hostedVersion with v3 will enable new version of payment page (for HOSTED integration scenarios) designed to support 3DS2 payment flows.
• style.resolution when set to 3x will return logos with 96 px of height for devices with higher PPI. Other possible values are 1x (32px) and 2x (64px).
• callback.returnUrl is the URL the customer should be redirected to, after the payment was successful, typically on the merchant's site.
• callback.cancelUrl is where the customer should be redirected to if he or she cancels on a redirect page (such as PayPal), or a payment attempt was denied and is not recommended to retry (Interaction Code ABORT).
• callback.notificationUrl is your notification endpoint where our system will send payment status updates to.

The full data model with all possible parameters can be found in our API Reference.

Read about the LIST Response if you want to render the payment page yourself.

You can specify the integration scenario in the LIST Request, using the parameter integration. The selected scenario will impact the operation URLs in the response object. If no integration parameter is specified, the LIST Request will default to Display Native.

These are valid parameters:

• HOSTED (Hosted)
• SELECTIVE_NATIVE (Selective Native)
• DISPLAY_NATIVE (Display Native, default. Legacy value: NATIVE_WITHOUT_PCI)
• PURE_NATIVE (Pure Native. Legacy value: NATIVE_WITH_PCI)

• URL: https://api.sandbox.oscato.com/api/lists
• Method: POST
• Authentication:
  • Login Name: registered company code
  • Password: generated API token

After a successful CHARGE is executed on a LIST object, the LIST will no longer be valid. This means that any other CHARGE on it will be declined, with the Interaction Code ABORT and Reason DUPLICATE_OPERATION. This is to protect customers and merchants from accidentally executing duplicate payments.

To provide a good customer experience, however, you should detect that case (via Interaction Codes passed as GET query parameters) on the page the customer is redirected to (indicated by the cancelUrl) and inform the customer that the payment was successful and that only the repetition was declined. Otherwise the customer may think the payment failed.

With the preselection sub-parameters, you can dynamically influence the LIST Response as follows:

• deferral: enables Deferred Payments or Payouts
• networks: Filters the response to:
  • REGISTERED: only accounts that were previously registered for the current customer
  • APPLICABLE: only networks that are to be entered by the customer, not registered accounts.
  • ANY (default): Both kinds
• direction: Specifies the kind of transaction, namely:
  • CHARGE (default): A payment transaction with a subsequent CHARGE Request
  • PAYOUT: A Payout to the customer

Re-using a LIST Session

A valid LIST can be seen as a payment session of the user. To enable payment conversion tracking through the Orchestration Platform, and to benefit from dynamic reactions of the payment page, we advise that you do not issue LIST Requests "blindly" but use existing LIST Objects if possible.

The object is a resource hosted by OPG. It is identified by the LIST's longId and can be requested any time again via HTTP GET like this:

https://api.sandbox.oscato.com/api/lists/[longId of LIST]

A LIST object remains valid for executing charges until either:

• a successful or pending (!) CHARGE through this list was executed, or
• 30 minutes have passed (this will also trigger a Status Notification). Note that in this case the timer starts counting from 0 again with any CHARGE attempt.

After that, the LIST object will become invalid for security reasons and no further CHARGE will be possible. In Display Native integrations, a GET on an expired LIST resource will return an HTTP 404 (not found) error. In Pure Native integrations, the object can be retrieved but its Status Code will be expired.

What this means for your implementation:

Use an existing LIST resource while it is still valid. In cases where your customer reloads the payment page, do not issue a new LIST Request, re-use the existing one. Only this will enable proper conversion tracking on your payment checkout through the Orchestration Platform. You can double-check the validity through a GET request on the LIST resource (see above) and/or listen to "expired" notifications. Only issue a new LIST Request if it has expired.

For some business cases it may be necessary to change the information provided in the initial LIST Request, for example if customers change the products in their shopping cart after they have been on the payment page but not yet completed the payment. It means that for the same "user session", an update of an existing LIST resource is required. For a new "user session", generate a new LIST, as described above.

The update is performed REST-style by sending an HTTP PUT to an existing LIST resource. The data structures payment, products and customer can be replaced as a whole. Other mandatory fields still have to be provided but cannot be changed.

• URL: https://api.sandbox.oscato.com/api/lists/ce3eec14-e186-4b7c-94a1-fde802412e69
• Method: PUT
• Authentication:
  • Login Name: DEMO
  • Password: BAtXa7yubR8ipuT

Note: LIST session updates can only be performed from the server-side system ("backend") and it needs your payment authentication credentials. After performing a LIST update, you may need to reflect resulting changes to the customer client (eg browser). Therefore, the LIST result should be read by the client again by passing it on from your server system, or by retrieving the LIST object direct from the Payment Gateway again, and showing the updated payment page. Also notice that the JSON structure is basically the same for a LIST request or update.

A LIST also holds information indicating which of the methods should be checked when the payment page is loaded. In the LIST Response, each object in networks.applicable and accounts, representing a payment method ("network"), has a boolean flag selected. If set to true this method should be displayed with a checked radio button and expanded form. The Orchestration Platform's Hosted payment page and AJAX library respect this value.

You can modify the flag by sending a PUT request to the method resource. Its URL is given explicitly in the self attribute within the method object.

Example request for default checking:

• URL: https://api.sandbox.oscato.com/api/lists/57bab64ee4b03ec79899332el/VISA
• Method: PUT
• Authentication: Not required (except for Pure Native integration)

As a consequence the selected property of all other methods in the list will become false. Note: other properties of a method cannot be changed through this API.

The selected attributes of all methods will be false by default, and we advise you to preselect the first method, as our AJAX library and Hosted Payment Page do.

Extended LIST Request

Sometimes you can improve the checkout experience by making more parameters available to the customer. For example, the products list could have more details about the goods ordered; or the customer might need to clearly specifiy their address. Some information may be required for risk management -- some invoice providers may restrict their offers if billing and shipping addresses are different, for example.

At the moment, the OPG accepts a long list of parameters in the LIST request, which you can find in our API Reference in the Server Payment API dropdown.

The given products will be shown to the customer in some networks, eg PayPal and Invoices.

product.amount reflects the price with respect to quantity. Therefore a quantity of 2 and amount of 25.00 means that a single product cost 12.50.

quantity defaults to 1.

The parameter clientInfo is optional, as is every sub-attribute within. But, if you provide any part of the object but not userAgent or acceptHeader, the following defaults will be used for these attributes if empty:

• userAgent: Mozilla/5.0
• acceptHeader: text/html,application/xhtml+xml,application/xml;q=0.9,/;q=0.8

These are widely accepted values. But just in case a PSP uses them for additional logic like risk management, we recommend that you set these attributes to the real values from the customer's original HTTP request.

While customer.number is the only required parameter, you should also consider customer.email. This enables basic identification that some providers use for risk management.

Phone numbers can be passed either by providing all three fields, countryCode, areaCode and subscriberNumber, or by providing a single unstructuredNumber.
You can provide an additional name structure in the billing section, but the general name defined as child of customer should be enough in most cases.

Addresses are specified better by the parameter addresses, which can cover shipping, billing, residential or other relevant address.

One interesting use case is where a customer wants different products to be delivered to different addresses. A merchant could assign certain products to be delivered to the customer's home, some to be delivered to the customer's office, and some even to be collected at a physical store. This scenario can be achieved by several parameters:

• A unique id for every customer address
• A list of additionalAddresses to specify several addresses with the same type, for example home and office shipping
• A merchantAddress (Boolean) to link one of the addresses as the physical store of the merchant
• A shippingAddressId to link each product item with its address

More LIST Use Cases

Delayed Payment Submission

The diagram shows how the user journey (on the left) relates to OPG API calls (on the right). Note how the CLOSING request from the Deferred CHARGE is separate from the Delayed Payment Submission.

Payment Account Validation

As soon as the customer submits their payment account data to the gateway (PRESET call), it can be validated for accuracy of format and also for existence of the account, availability of funds, and other risk checks. This means that if something goes wrong, the shop can give immediate feedback to the customer and offer alternative payment options to minimize user dropout.

This is configurable through the backend, so please contact us. The options are:

• NONE: No check will be made
• VALIDATION (default): An Orchestration Platform-internal check on formal validity of the account number will be made
• PROVIDER_CHECK: The internal verification (see above) will be made first. If this is positive, a check with the provider will be made, if available, that does not reserve the funds. For card providers, this means a zero-value authentication. For other methods it can mean a dedicated risk check with customer data. If no such provider check is available, none will be made and the transaction will be processed as usual.
• PREAUTHORIZATION: The internal verification (see above) will be made first. If this is positive, and the method and provider permits it, the corresponding funds will be reserved (preauthorized). If a reservation is not available, there will be a fallback to the non-reserving provider check as explained above. Again, if this is not available either, the transaction will be processed as usual.

LIST Request

As shown in the delayed payment submission diagram, the first step is to submit a LIST Request. For Delayed Payment Submission, an extra LIST parameter is needed: the optional Boolean value presetFirst. If this is set to true OPG will know that the payment submission should be preset, not charged direct. A summaryUrl must also be provided in the callback object if presetFirst is set to true. This way OPG can direct the customer to the desired summary page after the PRESET.

Delayed Payment Submission works with all Integration scenarios. After the LIST response is returned, follow the standard approach for rendering the payment methods, for example Native Integration, Selective Native, or it is rendered by the AJAX library.

PRESET Request

When the customer enters their details on the payment page and makes a submission, there should be a PRESET request (to the operation URL) using the customer's data.

In all integration scenarios except Pure Native, this will go direct to the OPG from the customer's client. It should contain the account data structure similar to the CHARGE request.

The AJAX library will handle this correctly on button click.

For a Selective Native integration without the AJAX library, the inter-frame communication must be triggered as described there, and an operation_response_event will be sent to reflect the server response.

The response to a successful PRESET call will show that the customer should be redirected to the summary page URL that was given in the corresponding LIST request. Again, the AJAX library will hande this. When implementing a Hosted or AJAX-dependent integration scenario, you might also want to receive, in the PRESET response, a full masked account data to show your customer which payment card has been used. In these cases, a simple GET on the LIST session can be done (GET on its self link); it will respond to the current session status and the preset account.

The operation URL of the returned object is where the subsequent CHARGE should be going to (see below).

Since the PRESET step also performs the configured validations if available (see above), the transaction could also be declined at this stage. A provider may, for example, decline to offer Invoice payments for this customer. This will be indicated by the Interaction Codes.

In most cases, the customer should then be returned to the payment page so that the existing LIST resource is reloaded and the remaining payment options presented again. In this example, Invoice may have been removed from the list, but the customer can still select another more secure method.

CHARGE Request

At this stage of the process, the customer is on the summary page generated by your shop system. Typically, this page would ask the customer to confirm the payment based on the summary presented. If the customer agrees, your server-side system should then make a CHARGE request using the operation URL returned in the PRESET response.

It looks like this: https://api.sandbox.oscato.com/api/lists/{LIST-longId}/charge

As the data has already been preset, no account information needs to be provided in the CHARGE Request body. In fact, only an empty JSON body {}  is provided as the request. This is also the reason why the CHARGE can go through your server-side system. There is no sensitive payment-data present here.

After the CHARGE call, your system should redirect the customer to the URL given in the response as redirect.url. In the success case, this will be the original success URL as specified in the LIST request (the thank-you page).

Payment Account Registration

There are 2 kinds of registration: Regular and Recurring.

Regular Account Registration

If customers return to your shop, they will see their previously registered payment accounts and do not need to enter their information again. This can increase conversion, especially with credit and debit cards or SEPA direct-debit bank accounts. For credit cards, the verification code (CVV, CVC, etc.) must still be entered by the customer because PCI DSS regulations prohibit the storage of this data. (If the payment provider in the background accepts this, as most acquirer interfaces do, this can also be omitted by using the Dynamic Verification Code feature).

The Dynamic Verification Code feature allows for hiding the CVC field for registered cards on our payment page. In this way,  you can submit a CIT payment without entering the CVC. Note that a payment provider needs to support this functionality before it can work on our end. Additionally, you need to use our payment page (so native integrations can't use the feature). 

In the 3DS 2 flow, the first payment will always require the CVC code. 

Recurring Registration

This kind of registration can be used to trigger Recurring Charges from your backend without customer interaction. This feature must be supported by the Payment Service Provider but does not require the verification code for any subsequent charges.

Customer Agreement on Payment Page

You can configure both types of registration. Since they need the customer's agreement, you should provide checkboxes in the payment page. If registration was enabled on the Orchestration Platform's side, you simply start a payment session with a regular LIST Request. In AJAX and Hosted integration scenarios, the payment page will then be rendered automatically and the CHARGE Request submitted correctly, so you won't need the following information.

If your system evaluates the LIST Response itself, there will be 2 attributes for each network: registration and recurrence, which correspond to the 2 registration types.

Depending on the configuration, these attributes can have the following values:

• NONE No checkbox should be displayed, the network is not registered
• OPTIONAL Checkbox should be displayed and can be checked by the user
• OPTIONAL_PRESELECTED Same as OPTIONAL, but checkbox should be pre-checked
• FORCED No checkbox should be displayed, but the account is registered anyway
• FORCED_DISPLAYED Same as FORCED, but a non-clickable and pre-checked checkbox should be displayed

Checkbox Rendering

In the OPTIONAL case, you can include an additional checkbox next to the network form, like this in the HTML source:

<input type="checkbox" id="allowRecurrenceVISA" name="allowRecurrence" value="true">
<label for="allowRecurrenceVISA">I authorize recurring charges on my account</label>

In the FORCED case, you can include an additional hidden value, as shown below, if you have a Display Native integration and need to pass the parameter from the client direct to OPG:

<input type="hidden" name="autoRegistration" value="true">

Checkboxes should look similar to those shown on the right. Note: a separate checkbox is needed for each payment network, because some networks do not support account registration and/or recurring charges.

Submitting the CHARGE

When the customer submits the payment data, the information indicating which checkboxes have been checked must be passed in the CHARGE Request. This is done through the parameters autoRegistration and allowRecurrence, which can have the values true or false (default). See the example.

Again, AJAX and Hosted integration scenarios already do this for you.

Receiving Registration Credentials

A CHARGE request like this will process the initial payment as usual. If the customer has not yet been registered, they will be registered at this stage together with the payment account, and the following Status Notification will be sent:

• Concerning entity = customer: This will be sent only once to your system with both attributes, customerRegistrationPassword and customerRegistrationId. These values should be stored on your side because they will enable you to use the payment accounts of this customer from within the Secure Storage again later.
• Concerning entity = account: This will contain payment network and accountRegistrationId which we will see again when manipulating individual registered payment accounts (as opposed to the whole customer registration). But in most cases you should not store this, because you will receive and use these IDs implicitly in the context of the use cases such as Account Update and Registration Page or Setting a Preferred Payment Account.
• Concerning entity = payment: This is the normal notification that only contains the customerRegistrationId which your system will receive for any payment by a registered customer.

Using a Regular Account Registration

If you enabled regular account registration (autoRegistration flag), the next time the customer does a checkout, you can submit the customerRegistrationId and -password with the next LIST Request as shown on the right (see customer.registration).

Using a Recurring Registration

As a result, you will get back not only networks.applicable in the LIST response, but also a list of accounts, which represent previously registered payment accounts. This contains only regular account registrations, not accounts that were (only) registered for recurring.

These payment accounts should be displayed above the selection of payment methods. The response will provide information such as a localizedForm and displayLabel to do this in the regular fashion.

Account Update and Registration Page

• A property operationType, which exists at root level must be set to UPDATE
• When updating a customer's payment methods, the customer's registration details must be included to identify the customer in question

In the pane on the right, there is a sample LIST Request for this operation.

• URL: https://api.sandbox.oscato.com/api/lists or https://api.live.oscato.com/api/lists
• Method: POST
• Authentication: Merchant Code and Token are needed

Registered Accounts: Only accounts that are allowed to be updated will be returned. If the account cannot be updated, it will not be shown. Where the payment account is updatable, the operationUrl returned in the LIST Response will point to an "update" instead of "charge" endpoint (for backwards compatibility with the existing integrations), eg "operation": "https://api.sandbox.oscato.com/pci/v1/5832ec78e4b0b0d86234c04bl/accounts/5832ebe41652164c14b86719a/update"

The associated formUrl / localizedFormUrl points to a form snippet with an expiration date and verification code, or there will be equivalent iFrame links for Selective Native.

Applicable Networks: Only payment networks with standalone registration capability will be shown. Their operationUrl points to a "register" instead of "charge" endpoint, eg https://api.sandbox.oscato.com/pci/v1/5832ec78e4b0b0d86234c04bl/SEPADD/register

The attribute operationType at the top level of the response object enables the client to see that this is an Account Update Request and then handle any desired follow-on action, such as displaying delete links for saved accounts.

The possible values of the operationType property are:

• CHARGE
• PRESET
• PAYOUT
• UPDATE

To demonstrate the Update LIST process, we have split the examples into different documents based on integration type:

• Native integration uses the concept of a localised form to display the data
• Selective Native integrations will retrieve and display an iFrame instead of a localized form, but the approach is similar
• Hosted Integration will handle the Update LIST on your behalf and while it is the easiest to integrate with, it offers the least control.

In an Update LIST, a customer can also set one of their registered accounts as preferred for future use. See Set preferred payment account.

Native integration requires some extra steps to update a LIST. Below, we show you how to:

• Register a new payment method for an existing customer

• Delete an existing registered network

Please follow the code examples in the pane on the right.

Step 1. Make the LIST Request passing the parameters operationType = CHARGE and account registration (id and password) object.

Step 2. Read the operation links from the applicable networks section of the response.
The operation link is the location of the POST Request that will register the new method. For convenience, a link to a localised form snippet will be provided. This snippet can be presented to the customer so they can enter their new details.

If another LIST Request is repeated with the same body as in Step 1, the newly registered card will be visible in the accounts and can be presented to the customer at the frontend. Hosted and AJAX integrations do this automatically.

Customer Score Dynamics

3D secure
In its standard setting, a customerScore value below 300 enforces a 3D secure check, a value above 700 disables it, and values in between let the provider decide in accordance with agreed defaults.

CVV/CVC (Dynamic Verification)
For registered credit card accounts, a customerScore value above 500 will skip CVV/CVC check if the PSP supports it. Case supported, the LIST response will return an empty form snippet for the account, which means that you can charge the customer without requesting CVV/CVC input. This also means that if a registered form snippet is present in the LIST response, you must implement it to collect CVV/CVC input from your customer before charge. Please notice that the OPG will decide for you if a verification code is needed for the currently available network. This dynamic verification logic is rather complex and depends on factors such as routing, network support, and CVC used in previous registration.
In practice, if you follow the implement once principle and always integrate the form snippet as delivered by the OPG, the verification code collection will be completely automated without need for extra logic on your side. The form snippets will always define what customer data is needed when sending a CHARGE request.

Set Preferred Payment Account

Integration

To let your customers set their preferred account, your system must initialize an Update LIST with the parameter updateOnly: true. As in any other LIST Request, the request body can determine if recurring payment account registrations should be returned by setting the parameter channel to RECURRING. Otherwise normal account registrations will be returned.

In any case the LIST Response will show the respective registered accounts, as in the example below.

This list should be visualized for the user in the common way. See LIST Response, Update LIST. To make a payment account the preferred one, your system makes a simple POST request to the setpreferred endpoint of that account. The URL is the self link as given in the response plus an appended /setpreferred.

So the request for the example above would be:

• URL: https://api.sandbox.oscato.com/pci/v1/585ab060e4b0f1217e9f4b3alr54bifbfla76esesbesj4k7ot/accounts/585a98bae4b091f43e2a0295a/setpreferred
• Method: POST
• Authentication: Not required (protected by the LIST ID)
• Body: {}

Further details on the behavior of setpreferred

• If the account was registered for standard registration and for Recurring Charges, the preference will apply to both. This means that the next Recurring Charge request by that customer would trigger a payment with that account, and it would show up as a first entry in the payment page of the checkout flow.
• If the preferred payment account gets deleted for one type of registration, the preference will still stay for the other type of registration.
• If the preferred account is deleted, the account that was the preferred account before will be reset as the preferred account again. OPG-internally; the account  preferences are listed in the order that they were set as 'preferred'.
• If none of the registered accounts were ever set as a preferred account, then the list is organized by time of registration: the most recently registered account will be top of the list.
• If the customer adds a new payment account, the previous preferred payment account will remain. In other words, the new account does not automatically become the preferred account unless no preferred account had been set (see above).

Delete Payment Account

To get all of the customer's registered payment accounts, and enable delete, your system sends a LIST request with the additional "allowDelete": true parameter, as shown in the pane on the right.

Note:

• This feature is supported by Hosted and AJAX integrations that will automatically render corresponding buttons. See below.
• If you are using an Advanced integration scenario your system should trigger the account deletion in a second step after the LIST, similar to the description of the Account Update and Registration Page issuing HTTP DELETE requests.
• The account registration data registration.id and registration.password must be sent in the LIST request.
• The allowDelete parameter can also be used to switch off the possibility of account deletion on the Account Update and Registration Page by setting it to false. Otherwise account deletion will be available by default.

When using our Hosted page or AJAX library, every registered account will render an additional "Delete" button at their side, as shown below. Naturally, you can modify its style via your custom CSS stylesheet.

When a customer clicks the Delete button, the following dialog appears:

The button-label and user-messages are localized, so if you are using the payment page for a language supported by the Orchestration Platform, they will show correctly.

Secure Storage

Payout Page

The JavaScript library of the AJAX Integration can be used to render the payout page. This results in a flow similar to the payment checkout:

The main difference is that you provide the parameter preselection.direction with the value PAYOUT in the LIST Request. See example.

This shows that a LIST for PAYOUT (as opposed to CHARGE) is requested. The effect is that all forms and endpoints for further actions that are returned will be coined towards a payout.

Apart from that, the flow is the same. The subsequent PAYOUT Request is also structured as a CHARGE and will be handled by the JavaScript library, unless you chose a native integration.

After a LIST is initialized for payout using the parameter preselection.direction = PAYOUT, any payment method returned in the list response can be deployed for credit operations. The payout URL can be found in the operation parameter and it will have this structure:

• Sandbox: https://api.sandbox.oscato.com/api/lists/{listId}/{network}/payout
• Live: https://api.live.oscato.com/api/lists/{listId}/{network}/payout
• Basic Authentication: Merchant ID and Token

The credit is completed by sending a POST request to the operation URL, and sending the proper payment account in the JSON body. This process works in a similar way to a charge operation in a regular LIST request.

The LIST for PAYOUT can also be requested in deferred mode, meaning that the customer submission of account details (the PAYOUT Request) does not yet trigger the money transfer. Instead, a consecutive CLOSING Request is needed, typically issued by the Merchant system after a staff member has reviewed the transaction. At this point, the transaction will be made through the configured providers.

For technical information on how to initialize and close deferred transactions, see Deferred Payments. The only difference for payouts is that there will be no reservation of funds on the payer's side.

Routing strategy options

On top of the standard routing configuration, you can configure routingStrategyOptions. It allows you to tune routing behaviour of transactions with stored credentials. 

Routing strategy options

The following routingStrategyOptions values are available:

• PREFER_TOKENIZED - execute transaction using only those providers where this card is already tokenized. That meansthat the provider stored this card in the provider vault and OPG is using the provided PSP token in a request. 
• PREFER_UNTOKENIZED - first use those routes which have this card tokenized. If no untokenized routes are available, use normal routing
• TOKEN_IRRELEVANT - use the first route available in routing independent of the tokenization status.

Default values

The default value for WEB_ORDER transactions for a registered account is TOKEN_IRRELEVANT. 

The default value for recurring transactions is PREFER_TOKENIZED.

For example, your configuration might look the following:

{...
 "routingStrategy" : "LEAST_COST",
 "routingStrategyOptions" :
 {
 "registration" : "TOKEN_IRRELEVANT",
 "recurring" : "PREFER_TOKENIZED"
 }
}

Examples

If you have the following routes: Route 1 (token available), Route 2, Route 3 (token available), then OPG will apply the following order based on the selected routing strategy:
 

If you have the following routes:Route 1, Route 2(token available), Route 3, then the following order will be applied:

AJAX Library Integration

Including the JavaScript Library

The Orchestration PlatformJavaScript Library (as known as AJAX Library) is used with Selective Native and Display Native integration scenarios to render payment components returned by API responses into your checkout page. With the library you can integrate a payment page in these scenatios with much less effort, since it implements all frontend payment features and only a single call is necessary to initialize it.

The most recent AJAX Integration package can be downloaded here: v3-3.28.0 (~2.5MB).

The zip file contains the following folders and files:

• build: The completely built and ready-to-use JavaScript and CSS files. The files should be put on the same folder level as your checkout HTML pages.
• example: An example HTML file in two versions, one linking the minified resources (JS & CSS) the other the development versions.
• node_modules: Node (npm) dependencies required to build the AJAX library.
• src: development source (based on node.js with the browserify addon).
• package.json: list of build and runtime dependencies, and development scripts.
• README.md: A starting point for developers (node.js experience recommended).

The main folder contains these files:

The example folder contains various implementation examples to help you getting started with our library.

You can integrate the respective JavaScript and CSS files into the header of an HTML page like demonstrated on the right pane. Make sure the HTML header contains the tag <meta charset="UTF-8"> for correct loading of the payment page.

Note that you also have to include the jQuery JavaScript framework yourself. If you want to support legacy browsers like Internet Explorer 6-8, Opera 12.1x or Safari 5.1+, we recommend a 1.12.x version for maximum compatibility (the latest 1.x version at the time of writing is 1.12.4).

We also recommend to set the viewport meta tag to enable the responsive UX, see example on the right.

The LIST Request needs to be initiated from your backend, because it needs to contain your sensitive OPG credentials. These should never be processed through the customer's browser. Read about this call in our LIST Request article.

As a result a LIST object will be generated in the OPG. This object contains all information to generate the payment checkout user interface (i.e., payment page). The only important part of the LIST Response is the link to this object (containing its long ID) or its actual longId attribute.

On the right pane you can find an example LIST Response with the link to the object >>

In order to have the payment checkout forms automatically injected into the current HTML DOM, you have to initialize the library.

You do so by specifying the HTML element you want to inject into first. Typically this is a <div> element. In the example it has the Id paymentNetworks and is referenced by jQuery syntax $('#paymentNetworks'). Then you trigger the checkoutList call from there.

An example call looks like this >>

When calling checkoutList you can provide additional options to configure and tune the payment checkout, listed below:

If the AJAX Integration library is correctly initialized and provided with a valid listId it will fetch all resources needed to build the payment forms from OPG automatically and inject it into the current HTML DOM.
Note that a generated LIST object expires after 30 minutes. Also it will only allow one successful Charge. See LIST Request for details.

Some hints to finalize your payment checkout page:

• To adapt the styling to your needs, we suggest to edit the (un-minified) CSS stylesheet provided in the beginning of this section.
• Make sure the provided icon hint.png is properly referenced and shows as a question mark icon next to the verification code input of credit cards, as well as the other icons in the desktop variant.
• Use the development version of the library, which is available for enterprise merchants, to do your own advanced customizations of behavior and generated markup structure.
• Further Payment Flow
• After the users' choice of payment method is complete they will be automatically redirected (returnUrl / cancelUrl / external PSP page) or get a response from your custom functions. In the case the browser goes to an external page for payment, they will still in the end be directed to the given returnUrl or cancelUrl.

Your backend will get a secure asynchronous notification about the current transaction status, which your system should listen to. See Status Notification for additional information.

If you wish to implement natively the Orchestration Platform's localized forms snippets returned from the original LIST Response, please refer to the Native Form Rendering article for details.

Provider Buttons

During a regular payment flow, the Orchestration Platform widget builds a list of payment methods and listens to a button-submit to process the next step (for example, capture a payment). This button is designed and implemented by the merchant, so to make this possible, it must have the ID submitBtn, which is linked to the AJAX initialization function checkoutList when building the payment page. However, providers such as PayPal impose the use of their own payment button, which has specific styling elements that cannot be changed by merchants. If you include such a provider in the list of available networks, you have the additional challenge of switching between a merchant's and a provider's payment buttons.

To solve this challenge, the Orchestration Platform AJAX library replaces this button on demand: if a user selects a provider like PayPal, its proprietary button is loaded. If any other method is clicked, the merchant button is loaded. This is done by defining a <div> with ID submitButtonContainer which will be used as container to submitBtn and the provider's button. The checkoutList function must be initialized with the parameter payButtonContainer pointing to this container defined by the merchant, as seen in the second example in the pane on the right.

After these initialization steps, the payment page will look like the example below:

Left: Submit button for other payment networks. Right: PayPal proprietary button.

From an integration point of view, this is the only change needed to start using PayPal checkout. On the Provider Contract configuration side, a new parameter JAVASCRIPT_INTEGRATION should be set to true, but you can request this change direct from our support team if you wish.

Delayed Payment Submission (Summary Page)

The Orchestration Platform's Delayed Payment Submission offers a unique solution: payment data is collected and cached ('PRESET') so the merchant has the chance to build a summary of the order before triggering the next step. With payment services that provide their own libraries, the summary page presents another two-fold challenge: First, the placement of the payment button has to happen there; and second, the proper process communication between the Orchestration Platform and these third party libraries must be handled.

To overcome these obstacles, the Orchestration Platform AJAX widget (the same one used for the Payment Page) must also be included in the merchant Summary Page and properly initialized with a different set of parameters. Its current implementation supports PayPal checkout, Apple Pay and Google Pay, and future updates will cover new methods. As you saw in the payment page, the library makes use of the initialization function checkoutList, but now with different parameters and extra handler functions:

• summaryPage:true tells the library that the current context is a summary page
• summaryPageHandler: optional functions that can be implemented by the merchant for better control of the payment process
  • doOperation: after an end-user clicks the provider-specific payment button, it will be triggered to handle the CHARGE process request to the OPG.
  • onResult: handles result object after success or error sent back from the OPG.

See the pane on the right for examples of these functions.

Summary Page Handler Functions

The idea behind the summary-page handler functions is to give merchants the opportunity to intervene in the payment process, by implementing their own way of handling the trigger and result processing of a transaction. In this way, you can introduce custom logic that starts other parallel processes in your systems; and you get better control of the payment flow without needing to rely only on the functionality provided by the Orchestration Platform. For example, with doOperation a merchant can send a signal to its backend saying that an item is being taken from stock. With onResult, you can define custom behavior of your user-interface based on the possible Interaction Codes sent by the OPG.

PayPal exception handling

There are some exceptional scenarios in which the merchant needs to cancel the payment operation mid-transaction. For example, it's possible that after triggering a CHARGE via doOperation a merchant sees that the item being ordered is no longer available. Ideally, the payment operation would be stopped before the end-user executes the payment on PayPal, meaning the PayPal popup should be closed by a trigger from the merchant systems. Since the introduction of PayPal checkout.js, such control is expected to happen in the frontend. To make this happen, the library has an exception handler that closes the popup. In practice, a merchant can trigger a popup-close by throwing a Javascript exception: throw new error.

For example, a merchant backend system can return an error to the frontend, which would then throw an error during doOperation execution to close the popup.

Display & Pure Native Integration

LIST Response

If you use Pure Native integration, the response will contain a few extra attributes -- see below. The idea is that this advanced data will be handled only by the merchant's server systems, whereas the Display Native data (above) could safely be passed to a browser for rendering the page in the frontend.

We will provide all the resources needed for display and charge of every payment network. With these resources you can easily generate a payment page of your own, the one the customer sees, within the frame of your corporate design.

The important links are:

• localizedForm: Links to an HTML form snippet detailing what information is needed from the customer to charge this network (eg credit card number). The language of the labels is dictated by the LIST Request's country or style.language parameters. The input fields depend heavily on the network. Their name attributes already correspond to the parameter names needed for the consecutive CHARGE. Only the ${formId} placeholder in "id" attributes needs to be replaced before displaying the form -- to avoid conflict with existing elements on the page.
• operation: The information entered on the form should be submitted to this OPG URL. It differs from network to network. Note: For Pure Native scenarios, you would submit to your own server first, and from there to the OPG. For Display Native scenarios, the submission comes directly from the browser. We recommend an AJAX request instead of form submission for this.
• logo: URL of a logo that can be displayed for this network
• validation: The custom validation service we provide for this network. The form data can be posted here asynchronously to receive live feedback on whether the information is considered valid.

You create the payment checkout page (most likely but not necessarily in HTML) dynamically and per transaction, with this data. Do not hard-code payment-network forms to keep the power of the 'implement once' principle. When the page is built, deliver it to the client (ie the browser), which displays it to your customer.

Visualization (red are parameters or resources to fetch; green are calls to execute):

Native Form Rendering

In Native intregrations, and in non-PCI applicable payment networks in Selective Native integrations, the LIST Responses will contain localized form snippets that can be used to integrate into your payment page.

Below is an overview of how to do this for the Native and Selective Native integration scenarios.

To retrieve the relevant localised form snippets, your system will make a LIST request specifying the integration type and country. See the example on the right.

Developers who have chosen a Native or Selective Native integration will have total control over where they present the form snippets. What follows serves only as a demonstration of how to do this; you are free to implement this at the frontend as you wish.
After retrieving the LIST Response, the associated payment networks should be presented at the frontend programmatically so that methods can be added or removed on the merchant configuration portal without the need for changes in the payment frontend of the integrated site.

In this simple example, all payment methods in a radio button list are displayed. When the radio button is selected, some JavaScript will fire and the associated paymentInputPlaceholder div containing the localised form snippet will be displayed. Some information from the LIST Response data values in these HTML controls are also populated; this is so they can be accessed later in the process. In this example, a Pay button is used for triggering the selected payment request; it is placed beneath the radio buttons.

As noted earlier, localised form snippets contain placeholders to identify the associated form. Before these snippets can be properly used, these snippets must be replaced with the relevant value.
To do this, inject the name of the payment method into the localised form at the frontend. Although there are different approaches for doing this, in this example you simply place all localised forms inside a div named ‘optileHtmlSnippet' then on the document load calling function, replace all form name placeholders with the form name. You can find the function opTemplateEngine we used in this example in the JavaScript file ‘template-engine.min.js

After this step is applied, the form snippet will now have the placeholders replaced. See the example below for SepaDD.

Before:

	<div id="${formId}-row1" class="row row1">

	   <div class="col col1"><label for="${formId}-iban">IBAN</label></div>

	   <div class="col col2"><input id="${formId}-iban" type="text" name="iban" class="text iban" autocomplete="off"></div>

	   <div class="col col3"><span id="${formId}-iban-message" class="message"></span></div>

	</div>

After:

	<div id="SEPADD-row1" class="row row1">

	   <div class="col col1"><label for="SEPADD-iban">IBAN</label></div>

	   <div class="col col2"><input id="SEPADD-iban" type="text" name="iban" class="text iban" autocomplete="off"></div>

	   <div class="col col3"><span id="SEPADD-iban-message" class="message"></span></div>

	</div>

If you wish to display checkboxes that let the user specify the payment method to be saved for future use, you must render this at the frontend, outside the form snippet.
In this example, a checkbox is rendered to handle this. Before rendering the checkbox and setting the desired value, the original LIST Response should be checked for each payment method to see if registration is supported. This is because some payment methods do not support registration.

The value in the LIST Response that illustrates this is the Registration property of the applicable network.

The possible values of this registration value can be seen in the console API but can also be summarised as:

• NONE – Registration is not supported
• OPTIONAL – Registration may be selected
• FORCED – Registration must be selected but not displayed
• OPTIONAL_PRESELECTED – Registration is optional but defaults to selected
• FORCED_DISPLAYED – Registration must be selected and displayed

Some logic must be in place in the frontend, to decide if the checkbox is rendered or not.

This information can then be retrieved on the Charge and included as part of the CHARGE Request.

This uses the same approach as above, but uses instead the recurrence property in the LIST Response to drive the logic.

There was a validation link in the original LIST Response for each applicable payment method. In this example, this information is loaded as a data value attribute in the snippet div.

On form submit, a jQuery function is needed to perform validation and should be implemented if the selected payment method is not an iFrame. Refer to Selective Native for iFrame validation in a Selective Native integration.

In this example, the implementation of the localised form validation is simply an Ajax call to validate the information contained in the selected form snippet. If there are errors, they will be rendered at the appropriate place in the form. Again, please note this example is for illustration only; you are free to implement this any way you choose.

Provider Buttons

Payment page rendering

Whichever integration scenario you have chosen, your system starts with the LIST request. If you are using the Delayed Payment Submission flow with a Summary Page before purchase, you can find out more under Delayed Payment Submission (this is easier to integrate because you don't need a proprietary button on the Payment Page). In other cases, your system renders the Payment Page by using the building components taken from the LIST response and should place the (regular) payment button in a dedicated container (eg <div>). When the user selects one of these proprietary "frontend" methods, your system should call a generic JavaScript code from the Orchestration Platform (see below) which will place the proprietary provider button into this container. (If you are using the full AJAX library, this container corresponds to the one given as the payButtonContainer parameter in the checkoutList initialization call). At the same time, your system should hide the normal payment button. The user will probably notice that the button has been replaced, but this is unavoidable because these kinds of providers don't let you use your own generic pay button.

JavaScript initialization on method selection

Your system now needs to know if it should call a JavaScript function when the user selects a method. For this, it looks into the HTML form snippet that was returned by the LIST response for this method or registered account. In any event, the form snippets should be integrated into your payment page, in accordance with our dynamic payment page rendering standards. Keep in mind that the form may not be visible to the customer (eg for PayPal) but it should still be loaded. If this snippet contains a <fieldset> tag with the attribute initcustomsubmitbtn then this gives you the JS function name (as a string) to call.

You can easily find the tag with this attribute (if it exists) with a CSS selector like fieldset[initcustomsubmitbtn]. When used in a JavaScript implementation that supports these selectors, this returns all fieldset elements that contain the attribute named initcustomsubmitbtn.

If there is no matching tag then there is no proprietary button for this method, and the regular payment button should be shown again.

If the function is given, it should be called as soon as the user selects the corresponding method. In other words, your system calls eval with the value of initcustomsubmitbtn and 5 parameters.

The 5 parameters are:

• networkObject: The JS object from the LIST response for the selected network. See the example.

• buttonContainer: a pointer to an HTML button container, as shown in the example on the right. Could also be a special "inner" container that only holds the proprietary buttons. See Button display below.
• paymentCallback: is called to handle the payment response. It should take only one argument which will contain the (CHARGE or PRESET) response from the OPG.
• operationUrl: is an optional string, if passed, it will be used instead of the url available in networkObject.
• listResponse: is an optional parameter containing extra information such as currency and amount.

Calling the initialization function will create the proprietary button and bind any subsequent click event on it. Naturally, it does not execute the payment at this point. This function acts as middleware between the payment page and the provider's proprietary JavaScript.

A click on the provider button will trigger the payment process as per the behavior defined in its own JavaScript code. The result will be given asynchronously to your given callback function (in the format of an OPG CHARGE or PRESET response) so that you can (typically) redirect the user accordingly.

On mobile devices, PayPal (for example) typically redirects immediately on mobile devices. On a desktop, the result is delivered once the user has completed the checkout in the provider popup. For your implementation, however, it makes no difference. It should just follow the redirect or react on issues (based on the Interaction Codes).

Button display

As mentioned above, your system should hide the normal payment button when the user selects a method that has its own proprietary button. We suggest you use CSS to make sure that all buttons appear at the same place so there is consistency for the user. However, they will be different, because they are provider-specific. Providers usually do not allow the buttons to be changed visually.

In a nutshell, every time the user selects a new method, your system must evaluate again which buttons to show or hide. As a general rule, the decision goes like this:

Every time the selected payment method changes, hide all buttons inside the container. Note that they cannot consist of a <button> tag but could simply be a <div> for example. Then either call the JS function to show the proprietary button, or show the generic pay button. You could also group all proprietary buttons into another (inner) container within the (outer) container (in which the regular payment button is located). This gives you clear separation between regular and proprietary buttons. In this case you would pass the inner container to the function given in initcustomsubmitbtn.

Delayed Payment Submission

You should include the Orchestration Platform's AJAX library in the Summary Page, as described under https://www.optile.io/opg#1842448 (again, look under Delayed Payment Submission). This will render any proprietary payment button, if required, or just handle the button click for you. In any case it leaves you with full rendering possibilities, so you should use it on the Summary Page even if you don't use it for Payment Page rendering. It does not make a difference at this point which integration scenario was used to render the Payment Page earlier as long as the LIST object contains a preset payment account.

CHARGE Request

You only need to implement the Frontend CHARGE Request yourself in Native Integration scenarios.

The customer's sensitive payment-account data originates from the customer's client. That's why we call it a Frontend CHARGE (as opposed to triggering a Recurring Charge which would be a Backend Charge). Under PCI regulations, the sensitive data takes different paths depending on the Native Integration scenario that's in place:

Display Native: The data is transferred from the payment checkout page in the customer's client (eg browser) direct to the OPG. Your server-side systems never touch the data and therefore don't have to be PCI-certified.

Pure Native: The data is transferred from the payment checkout page to your own server-side systems. From there it is submitted to OPG (server to server). This means your system may store the data itself. In the case of credit cards, your system must comply with some PCI DSS standards and you must be certified.

The endpoint for the CHARGE Request is given as operation attribute for the method in question, in the LIST Response.

• URL: https://api.sandbox.oscato.com/api/lists/ce3eec14-e186-4b7c-94a1-fde802412e69/VISA/charge
• Basic Authentication: ABC123:BAtXa7yubR8ipuT
• Method: POST
• Headers
  • Content Type: application/vnd.optile.payment.enterprise-v1-extensible+json
  • Accept: application/vnd.optile.payment.enterprise-v1-extensible+json

The CHARGE Response is returned to the system that issued the request. This is often the customer client (eg browser); in Pure Native scenarios it is typically the merchant backend. At the same time, a Status Notification is sent from the OPG to the merchant backend (as specified in the notificationUrl), containing the same status information. It can be used as a reliable source of status information for the merchant. On the other hand, the CHARGE reply should be evaluated in terms of feedback to the customer and how to direct them further. A CHARGE response can contain redirect information for methods that require the customer to finish a payment on their own platforms.

Some payment methods will require the customer to be redirected to their own web platform to authorize a purchase. Many alternative PSPs and especially 3D Secure methods rely on redirection, therefore it is important to understand how the CHARGE response represents them.

In such cases, a CHARGE can be considered a two-step process. It will initially stay in pending state and will only move to a successful (or failed) state after the end customer finishes interacting with it. From an integration perspective it means you should evaluate first the Interaction Code and Reason of the response and parse the redirect URL accordingly.

A positive CHARGE response of redirect methods will initially be in pending state, which is represented by status.code = pending and interaction.reason = PENDING. Its redirect object will contain parameters to inform how your systems should handle the redirection as follows:

• url: the base URL to which the request should be sent to.
• method: the request method used. It can be either a POST or a GET.
• parameters: an array of parameters to be sent with the request. In case of a POST redirect, the parameters should be sent as in a form submission with header Content-type = application/x-www-form-urlencoded, in a GET redirect the parameters should be attached to the main URL as query string elements. If account placeholders are used (see next item on this list) the following key-value pairs can be present and should be replaced by the respective input form data collected from the end user.
  • holderName: ${account.holderName}
  • cardNumber: ${account.number}
  • expiryMonth: ${account.expiryMonth}
  • expiryYear: ${account.expiryYear}
  • verificationCode: ${account.verificationCode} 
• containsAccountPlaceholders: a boolean that indicates if any of the parameters above contain placeholders for payment account data. Some PSPs will expose payment data in full details as redirect parameters, so to enable merchants that have PCI certification SAQ-A to still process such redirects, the account-related parameters will contain placeholders that our Ajax library replaces with input data from the frontend iFrames. For merchants that have are on higher levels of PCI certification SAQ A-EP or SAQ-D, these placeholders need to be filled with data collected in the frontend through regular forms.

After the user completes the purchase through the redirect window, a regular CHARGE response containing either a success or error case will follow. Note that these responses also contain redirect URLs to the success or error pages defined by your shop, and should be parsed by the exact same method described above.

After a successful charge request, the response body will also contain a masked account used for the charge. This can be used for giving the end customer a better summary of the payment process. The data can be found inside the maskedAccount parameter in the response body. See the example in the pane on the right. 

It is important to note that when using client-side charges (via our AJAX library, for example) the masked account data will not be sent on the response.

A typical successful CHARGE response looks like this:

If the request is accepted and the Interaction Code in the response indicates PROCEED, the response will contain the redirect structure with information on where the user should be redirected. Only in Pure Native scenarios, where you did not provide the callbacks in the LIST Request, must you generate the redirect in your backend yourself.

There are two reasons for redirection after a successful request:

1. Payment was completed successfully. In this case the redirect will lead to the merchant's success page.
2. Payment was initiated but more information is needed from customer, eg 3D secure code for Visa or MasterCard, or authentication and confirmation at the site of a redirect network such as PayPal or Sofort. In this case, the redirect will lead to the acquirer or PSP website to continue the interaction with the customer.

Regardless of the reason, the task of your system is simply to redirect the customer's browser according to the information provided in the redirect section.

The key-value pairs provided in the parameters section should be attached as GET-Parameters to the given url. This enables the next page to display additional information about the payment status. For example, it may provide data for the remittance account in the case of invoice or prepaid payment networks (see Status Notifications for a list of possible keys relating to remittance accounts).

For Hosted Integrations in an iFrame especially, the suppressIFrame flag indicates whether the redirect should happen in the top frame of the browser (true), hence suppressing a potential iFrame. This is required for explicit redirect methods, such as PayPal, or redirects to one of your pages. It should therefore be the default, as it is in the Orchestration Platform's hosted page and AJAX library. Only if false is explicitly given should the redirect  be in the current frame. This can be useful for hosted white label pages by third-party providers.

The bottom response parameters timestamp, operation, resultCode, and identification are intended for enterprise merchants only; you may not need them.

If there was a potentially recoverable error during payment, the following Interaction Codes may appear in the response:

• TRY_OTHER_NETWORK
• TRY_OTHER_ACCOUNT
• RETRY

In this case, no redirect section is contained in the CHARGE response >>

In this case you should inform the customer of the failed payment. Depending on the Interaction Code the customer should be asked to select another payment method or provide correct account information.

• TRY_OTHER_NETWORK Re-read the LIST object and render the payment page again accordingly, because the used network is probably not available anymore.
• TRY_OTHER_ACCOUNT  A different payment account should be requested from the customer (it can be the same method, eg Visa credit card, but a different account number would be expected). The LIST Object has not changed.
• RETRY There is some other problem that may require updated account details, or to retry this later. The LIST Object has not changed.

When integrating our services natively, localized messages can be exhibited by mapping the interaction codes returned by OPG with the respective values provided by our live environment in a resource JSON file:

https://resources.sandbox.oscato.net/resource/lang/MERCHANT_CODE/language/checkout.json

In this URL, two parameters need to be adapted:

• MERCHANT_CODE: your Merchant Code
• language: a language code composed of ISO 639 two-letter language codes and ISO 3166 two-letter country codes separated by an underscore character. Similar to RFC 1766.

The resource file lists the interaction codes and reasons, with their related descriptions as specified in the table Possible Interaction Code Values, plus translation for other page elements such as button labels and titles. An extract of the file from https://resources.sandbox.oscato.net/resource/lang/DEMO_SK/en/checkout.json is shown in the right pane.

We are continuously working to provide more languages in these resource files, but where a specific language is still not available, OPG will automatically fall back to English.

For some methods with a multi-step flow, eg Installment, the Interaction Code RELOAD will be returned. As with TRY_OTHER_NETWORK the LIST should be reloaded in this case and displayed again. See ACTIVATE Request for background information.

In some cases (duplicate payment, system failure, fraud detection, session expiration, etc.) the response will advise to interrupt payment process via Interaction Code ABORT.

This will be coupled with a redirect to merchant's Cancel Page, if the cancelUrl parameter was given in the LIST Request.

As with successful payment processing, if a redirect section is found in the response, you should simply redirect the customer's browser accordingly.

If the reason was a duplicate payment (interaction reason DUPLICATE_OPERATION), ie a Charge on a List object after there was already a successful Charge on that List, we recommend that you add a hint for the customer on the Cancel Page saying that the payment was already done and that only the repetition failed. Otherwise the customer may think the entire payment was denied.

Validation API

The validation response contains user-friendly messages that should be displayed in front of the input fields, for example "Correct", "Invalid expiration year", "Missing card number" etc. The messages appear in the language that was specified in the LIST Request.

You can get the validation API endpoint from links object within the LIST Response. The endpoint is unique for each of the payment networks returned.

• URL: https://api.sandbox.oscato.com/pci/v1/0a15894e-732c-4afb-a6ec-2edd7be4edb8/DEMO/de_DE/VISA/standard/validate
• Method: POST
• Basic Authentication:
  • Login Name: Merchant Code
  • Password: Payment Token
• Headers:
  • Accept: application/vnd.optile.payment.simple-v1-extensible+json
  • Content-Type: application/vnd.optile.payment.simple-v1-extensible+json

The JSON parameters to be sent are equal to the input field names of the HTML payment form snippets linked in the LIST Response. For example, the validation request for a Visa credit card should have parameters that match the form snippet on the right pane (notice that the name attribute is equivalent to the parameter name to be sent in the JSON request body:

Commonly used parameter names are listed below (this list is not exhaustive and may change with new payment networks):

• holderName
• number
• expiryMonth
• expiryYear
• verificationCode
• bic
• iban

Validation passed

Empty field value or invalid input value:

HTTP 200 OK

Invalid format of JSON in the request:

HTTP 500

Activate Request

If a method uses the Activation flow and you do not rely on the Orchestration Platform's AJAX or Hosted Integration to take care of it, your implementation should detect this case as follows.

The initial handling of the payment page stays as is. The provided form will be shown to the customer, the entered data submitted to the operation endpoint given in the LIST Response.

If the method uses the Activation Flow you will notice that the endpoint ends in activate instead of charge, but you do not need to detect that in your implementation. The part your system should pay attention to will come in the response.

If the customer can proceed with the checkout, the response will be something like >>

The Interaction Code RELOAD tells your system to reload the LIST Object (via GET on the LIST Resource) and generate the payment page again with its content. The provided form of the activated network has now changed and will request additional information from the customer.

In order to not confuse the customer, the selected flag for this method will now be true, indicating that this method should be preselected and expanded when generating the payment page again.

From there the regular CHARGE process applies again: The form content gets submitted to the operation endpoint given in the LIST Object (which will now end in charge), and the response should be evaluated.

URL: https://api.sandbox.oscato.com/api/lists

URL: https://api.sandbox.oscato.com/pci/v1/582994c0e4b0b0d86230c524l/INSTALLMENT_SEPA/activate

Method: POST

URL: link.self (you find that link in the LIST Response)

Example: https://api.sandbox.oscato.com/pci/v1/582994c0e4b0b0d86230c524l

Method: GET which return the Installment Plans

URL: https://api.sandbox.oscato.com/pci/v1/582994c0e4b0b0d86230c524l/INSTALLMENT_SEPA/charge (use the operation URL from the Reloaded LIST in Step3)

Method: POST

CHARGE Request Example >>

If one chooses to implement Sepa Instalments in a Selective Native or Native implementation, then there is a small amount of effort required to render the process on the front end.

After the initial LIST Request is made with Selective Native or Native as the implementation parameter the LIST Response  will contain a series of applicable networks each with an associated localized form (or frame if it is selective native and a credit card network).  These html snippets (or iframes in selective native) should then be used to render the front end presented to the customer for each payment network.

These localised forms will contain {form-id} placeholders that need to be replaced with the network.code (highlighted in fig 1) value of the payment network. This can be accomplished easily using JavaScript. See the section 'Populating form snippets' in the Localized Form document for a simple approach to illustrate this. Also ‘network code' and ‘self' link should be saved as data attributes in the html element displaying the localised form as they will be needed later in the sepa-installment process.

After this replacement happens a form  asking the customer to enter his/her date of birth will be displayed. If the year of birth field is empty it means that the JavaScript in the html snippet didn't execute and you should recheck your placeholder injection.

On submit the front end should make a call to the operation url returned in the previous LIST Response (illustrated earlier in this article). Opposite is some sample javascript to illustate this submit process.

On the call-back the response should be checked to see if it was successful. If the response was successful it will prompt that the form should be reloaded. This can be accomplished by using the self link for the payment network returned in the original response (see example above).

If it was unsuccessful then this information can be displayed to the user.

After the reload ((*) in JavaScript sample opposite) the form will now be in an activated state and the operation from the new RELOAD response will be the end CHARGE URL (see sample JSON earlier in this section). This can be used for the end submit action for the next form request. The ACTIVATION (reload) response will also contain a series of instalments that need to be used to populate dropdown values in the new localised form html snippet.

Repeat the {form-id} replacement process to load the network code into the localised form snippet and additionally at this stage, because the activated SEPA-Instalment form needs the drop down of potential rates to be populated and handled, an additional piece of JavaScript is required ((**) in JavaScript sample opposite). This is accomplished by finding the name of the postCreate function in the form snippet that is associated with the localised snippet's fieldset. There is a placeholder for formData there that can be populated with the instalments returned by the last LIST Request ((***) in JavaScript sample opposite).

At this stage the completely rendered form snippet will be displayed on the frontend and the included JavaScript will handle the form changes.

On submittal of the above form it's data should be serialised and submitted as part of the CHARGE Request. The process will then be complete. See opposite for sample Javascript to handle this.

Skip sending CVC code

If you're switching to a provider that doesn't require a CVC code for registrations, you might want to configure that you don't have to send the CVC code. This will spare you the need to migrate the token and will make the contract switch easier. To help you achieve it, we have made it possible to configure the CVV/CVC validation (verificationCode processing) on a contract level.

This functionality is available only for native integrations.

Verification codes

Card verification code is a series of numbers (in addition to the bank card number) which is embossed or printed on the card. It is used as a security feature where a personal identification number (PIN) cannot be manually entered by the cardholder (as they would during point-of-sale or card- present transactions). Card verification code has different names with different providers: 

Configuration options

For some providers verification code is required for cardholder-present (WEB_ORDER or MOBILE_ORDER) transactions, or for transactions, where card is stored for later usage or when card is used for the first time. To be able to configure verification code validation on OPG side OPG has enabled three possible validation options (VerificationCodeOptions), configured on contract level.

These parameters are not available for the following networks: 

• DINERS
• MAESTRO
• CARTEBLEUE
• POSTEPAY
• VISAELECTRON

Scenarios

Selective Native Integration

If you implement this yourself, you should first be aware of the behavior of the LIST Response when it generates the payment page (client- or server-side).

For each method in accounts or networks.available that falls under the scope of PCI DSS (ie credit and debit cards), the LIST Response will not provide the links form and localizedForm. This means that your system will not be able to include the HTML snippets for the corresponding forms, even if you forget to change the implementation.

Instead, it will provide the URL to the iFrame in links.iFrame and the suggested height of the iFrame (in px) in the iFrameHeight attribute. So an object representing a payment method would look something like this:

Your system should place the iFrame into the payment page with these parameters >>

Web browsers will not allow direct access into the iFrames from the outside (the surrounding page). But in the standard version of Selective Native, the submit button is still outside, on the payment page. We therefore need predefined messages that can be exchanged between the frames.

Briefly, the essential messages are:

• validation_response_event Sent from the iFrame to the outer page to reflect whether there is valid account information entered so that the outer page can enable/disable the Pay button (for optimized user experience).
• payment_action Sent from the outer page to the iFrame as soon as the Pay button is pressed, telling the iFrame to do a final validation and submit the payment operation request.
• operation_response_event Returns the response of a payment operation request from the iFrame to the outer page.

To provide an excellent user experience, Live Validation of the data entered by the user should be enabled (in the iFrames it is enabled by default), and the Pay button should be disabled for as long as the entered data is not formally valid.

This means that your page should decide whether to enable or disable the button, every time the user selects a payment method and the payment page is generated. If the selected method has a plain form (without iFrame) and contains a <fieldset> element or it contains an iFrame form, disable the button for now. If there is data entered already, do a validation right away (see below). Otherwise, enable the button which is required for methods that do not require any input on this page (eg PayPal).

Your page should do a Live Validation of the entered data for methods with plain forms, for example when the user stops typing for a moment or leaves the input field. This is particularly important because your payment button will be disabled in the first place in these cases. If the response signals "valid": true then the payment button should be enabled, otherwise disabled (again) and the hints shown next to the corresponding input field(s).

Inter-Frame communication comes into play if the selected method is subject to PCI and therefore provides an iFrame instead of a plain embeddable form. By default, the iFrame will do Live Validation itself and display hints accordingly in the form. The iFrame will also send the result to the outer page in the form of an validation_response_event. Your page should listen to this message and enable/disable the payment button accordingly.

Here is some JavaScript / jQuery example code >>

Note that event.data["response"] contains the full data structure of the Live Validation Response. This means, for example, that event.data["response"]["valid"] corresponds to the valid attribute from the direct API response.

You can disable Live Validation of the iFrames by requesting them with a liveValidation=false query parameter. For example:

https://resources.sandbox.oscato.com/paymentpage/iframe.html?listId=57c03359e4b0c50a7e8007cfl&network=VISA&liveValidation=false

With Smart Switch your customers do not have to select their credit card brand, it will be detected automatically.

To enable this feature with Selective Native you open the iFrame of a credit card form with the added parameters smartSwitch=true and networks (plural, not network) with a list of credit cards codes (corresponding to the credit cards of your configuration). For example:

https://resources.sandbox.oscato.com/paymentpage/iframe.html?listId=57c03359e4b0c50a7e8007cfl&smartSwitch=true&networks=VISA,MASTERCARD,AMEX

An iFrame that is integrated like this will detect the credit card scheme of the entered number, during typing, and send the detection result to the parent frame in the form of a detection_response_event. It can be consumed as the other response events and always has the following structure >>

The method attribute will reflect the detected method from the list that was given as URL parameter. If none of those were detected, the value will be null.

The payment_action message should be sent from the outer page to the iFrame when the Pay button is pressed, telling it to do a final validation (using the same call as Live Validations). If final validation succeeds, submit the card data direct to the OPG for payment. At this point, you should disable the Pay button and, ideally, show a spinner to indicate that the user now needs to wait for feedback. When the response arrives, the result will be sent as a separate message from the iFrame to the outer page again.

If the validation fails, there will be a validation_response_event message as shown above. These situations should be very rare if you implemented the enabled/disabled  Pay button correctly. This is because the button will only be enabled if the validation succeeded already (see above), although it could still happen, eg due to race conditions. In any case there is nothing to do at this point because the Pay button should be disabled after the payment action, and stay disabled. Only the updated validation hints will be shown in the iFrame.

If the validation succeeds, the iFrame will execute the payment operation (eg CHARGE) and the response will be sent as operation_response_event message from the iFrame to the outer page. This means that on the page there should be a listener for this event, taking the response data and reloading or redirecting the user according to CHARGE Response.

Only one response event will be sent as a result of a payment_action.

Here is some JavaScript / jQuery example code you can use to get started. It is an extension of the above example code >>

For message type operation_response_event, event.data["response"] will hold the entire CHARGE Response. That means you can get the Interaction Code with event.data["response"]["interaction"]["code"], or the potential redirect URL with event.data["response"]["redirect"]["url"] and so forth.

With this data at hand, all that is left to do is react to the response, as described under CHARGE Request.

Specific Actions:

Besides the payment_action message, the following messages can also be sent to the iFrame:

• validation_action Will trigger an explicit Validation of the payment account data currently present in the form and will cause a validation_response_event as soon as the validation was completed (see above).
• operation_action Will trigger the payment operation directly (i.e., without extra validation) with the currently entered payment account data and this will cause an operation_response_event as soon as the operation was completed (see above). This could be used instead of the payment_action if the iFrame does Live Validation in order to avoid the delay of one extra validation.

Since the payment_action we have described combines these two actions, two different response events can result.

In principle, all requests are subject to race conditions. Therefore your page can assign IDs to actions that will be returned with the corresponding response event. This way your implementation could detect messages that overtook other messages.

You can set and read this ID in the attribute action_id, so you could:

selected_iframe.postMessage({"type": "validation_action", "action_id": "42"}, opg_origin);

You can access the value in the response event through event.data["action_id"] and thereby match the response to its original request.

If your implementation does not make repeated action requests, this would not be neccessary.

The frame snippets presented to the user can be styled and translated. To do this, simply pass in the desired language and cssOverride URL in the style attribute in the original LIST Request. Afterwards, the frame will have its contents translated and style applied. See the screenshot below for a simple example of where the frame has been translated to French and has been styled to a dark gray background and heavy bold text.

If you wish to use Selective Native to integrate other payments methods (other than credit cards) then you could use the Orchestration Platform's localized forms snippets returned from the original LIST Response. Find out more about implemeting these localized forms.

Because the frame that handles payment does not contain checkboxes for registering payments or allowing an input payment method to be used in recurring payments, these must be handled outside the frame.

The checkbox must be rendered after the frame. Then, on submit, the values of these checkboxes can be used in the frame's post message request to specify whether this payment method is to be registered or used for recurring charges. In the example opposite the values allowRecurringPayment and allowPaymentRegistration will have been retrieved from the checkboxes via JavaScript.

For a broader overview, go to displaying registration checkboxes in Localized Forms. 

Client-Side Encryption

The integration packages are available as zip files here:

• op-client-side-encryption1.0.6.zip (smaller file size, compatible with modern browsers)
• op-client-side-encryption1.1.0.zip (bigger file size, but providing fallback support for legacy browsers IE 9 and 10, and Android 4.x.x)

Inside each file, next to the root files with development meta information, you will find these folders:

• demo: an html file you can view locally to interactively test the encryption of custom account data with public keys (see below). There is no server/OPG interaction here.
• dist: the actual distributable encryption library in ready-to-use minified format:
  • op-client-side-encryption1.x.y.min.js The versioned filename which we recommend to use in order to easily keep track of version changes.
  • op-client-side-encryption.min.js A copy without versioning in its file name to represent the latest version. Could be used for on-the-fly upgrades of the library, which requires more careful management.
• src: the development source files

The integration starts with a LIST Request, but with CSE you must use the same integration parameter as for Pure Native integration:

"integration": "PURE_NATIVE"

If you submit encrypted account data using any other integration scenario (including the default), you will get this error: "Unrecognized JSON property 'encryptedAccount'…"

After the user has entered their payment-account data, the web client is expected to build a JSON object from it and then encrypt it. The attribute names contained in this object are the same as the input field names used in the equivalent HTML form from the Orchestration Platform's form-based API. In other words, the object should have the same attributes as those in the Orchestration Platformaccount object used for CHARGE Requests.

The most common attributes for Charges are:

• number
• holderName
• verificationCode
• expiryMonth
• expiryYear
• IBAN
• BIC

This payload should be represented as shown on the right pane.

The minified JavaScript encryption library can be included in your web page, either by loading it direct in a <script> tag, or through require or import. In technical terms, the library provides an asynchronous API to encrypt user data in accordance with the standards below.

The library exports the function: opEncrypt(accountPayload, publicKey, [timestamp], [listId])

This function returns a Promise Object (also called a "thenable"), which resolves to the encrypted string through its callback. This string is the encrypted data structure that you can send through your server to the OPG. We chose this implementation path to make the library compatible for a wide variety of use-cases, including server-side use and advanced client-use scenarios that benefit from an asynchronous programming style.

• accountPayload: the JavaScript object (not a string) with any of the attributes available in the Orchestration Platform account object used for CHARGE Requests
• publicKey: the public key in JSON Web Key (JWK) format (as JavaScript object), which you already received from the Orchestration Platform
• timestamp (also known as generationTime - optional): the timestamp of the encryption. If it is not given, or given as null, the current client time is used as default and added to the payload. Input format is ISO standard, eg "2017-09-26T08:41:01.859Z"
• listId (optional): the longId of the OPG LIST session on which the Charge would be executed. As it is an optional parameter, listId will also not be included into the payload if it is not given.

The function returns a Promise instance which resolves to a string containing the JSON Web Encryption (JWE) encoded account data. This string is the data that can be passed on to the OPG.

Your server sends the payment-account data in an (authenticated) CHARGE Request to the OPG. The parameter to hold that data string is called encryptedAccount (which replaces the account object of a regular CHARGE Request). Note: registration flags are still given, unencrypted.

The LIST session for this CHARGE transaction needs to correspond to the listId used during encryption. For security, the encrypted card data will not work with any other LIST sessions. The response from the OPG will be the same as with an unencrypted CHARGE Request.

Never send unencrypted card data to your server systems.

If you want to build your own encryption client, you will need these extra details. You don't need these details if you simply want to use the JavaScript encryption library provided by the Orchestration Platform.

The encrypted content is transmitted as a RFC7516 JWE (JSON Web Encryption) data structure in JWE Compact Serialization format. All of the user's input (see above) is encoded in key/value pairs in the encrypted JWE JSON payload. The Orchestration Platformonly supports AES256-GCM with RSA-OAEP as the encryption scheme. Therefore, the JOSE header will always look like this:

{ "alg": "RSA-OAEP", "enc": "A256GCM" }

It's base64 representation will accordingly always be:

eyJhbGciOiJSU0EtT0FFUCIsImVuYyI6IkEyNTZHQ00ifQ

As specified by RFC7516...

• the JOSE header
• the RSA-OAEP encrypted symmetric content encryption key
• the randomly chosen encryption initialization vector
• the AES256-GCM encrypted payload JSON object literal with the above JOSE header passed into the encryption operation as "additional authenticated data"
• and the resulting AES256-GCM authentication tag output…

…are each base64-encoded and concatenated using a "." character as separator. The resulting string is the one that must be sent to your server and then passed on to the OPG.

Backend Use Cases

Refunds & Cancellations

Payout requests generally only need a payment data structure that includes the amount to be refunded. With some PSPs, you may also be able to submit an empty payout request translating to a full refund. But with other PSPs this does not work for partial refunds that have been issued beforehand—therefore, we do not recommend it. To stay on the safe side, always specify the amount to be refunded.

In some cases you may have to specify which products are refunded (eg if you use invoice payments). To do that, you can specify a products array after the payment structure, similar to the product information you provide in an extended LIST Request. It is important to note that most PSPs will give you an error if you try to refund products that were not specified during the initial checkout.

First partial refund example

Suppose the initial checkout Charge had an amount of 19.99 € and got the longId value 26d5b179-a3b5-4aa3-ad98-3074b41e7d2b. From there, you build the payout URL and post an amount less than the original amount (partial refund) as follows:

• URL: https://api.sandbox.oscato.com/api/charges/26d5b179-a3b5-4aa3-ad98-3074b41e7d2b/payout
• Basic Authentication: UBUY:BBtXa7yubR8ipuT
• Method: POST

We issue a second refund that covers the remainder of the Charge amount—after which the customer will have been fully refunded.

• URL: https://api.sandbox.oscato.com/api/charges/26d5b179-a3b5-4aa3-ad98-3074b41e7d2b/payout
• Basic Authentication: UBUY:BBtXa7yubR8ipuT
• Method: POST

If your PSP does not support partial or multiple refunds for a one payment transaction, you should receive an error like: "Multi-refund is not supported".

If your refund amount exceeds the amount originally charged, you will get an error from the OPG. To demonstrate, we issue a third "partial" refund on the same Charge, that was already fully refunded.

• URL: https://api.sandbox.oscato.com/api/charges/26d5b179-a3b5-4aa3-ad98-3074b41e7d2b/payout
• Basic Authentication: UBUY:BBtXa7yubR8ipuT
• Method: POST

Where you need to create an 'exceeding Refund' and your PSP supports it, the validation on the OPG should be turned off. Please contact us in this case.

Some PSPs ask you to list the products which are to be refunded. Make sure the product.code and amount (ie quantity) of the refunded products match those from the original CHARGE Request. If the Payment Gateway finds a product with the same code but the amount does not match, it will return an error.

• URL: https://api.sandbox.oscato.com/api/charges/26d5b179-a3b5-4aa3-ad98-3074b41e7d2b/payout
• Basic Authentication: UBUY:BBtXa7yubR8ipuT
• Method: POST

The products list behaves exactly as in the extended LIST Requests. The quantity is optional and defaults to 1, and the amount is the sum after considering the quantity.

The response and behavior in respect of partial ('multi') refunds is equivalent to those from simple Refund scenarios.

Cancellation

Refunds can be applied if a Charge has been processed by the provider. But there are situations where a Charge is queued on the provider side (or the Orchestration Platform side) before it gets processed. The reason for this is often a provider-internal batch processing schedule. In this situation, the provider typically allows you to cancel this transaction. In the Orchestration Platform's status space, such a transaction will have Status Code pending and the Status Reason processing_scheduled.

Canceling such a transaction means it will not be processed by the provider—so generally no processing fees will apply, and the end customer will not see any related transaction on their account statements (as opposed to refunds where a debit and a credit will typically be visible). In most cases, a cancellation can only be done completely. If you want to return a partial amount, you can wait until the Charge was processed and then do a partial refund instead.

To request the cancellation, send an HTTP DELETE request (without a body) to the CHARGE resource (as accessible by its links.self URL).

Example Request

• URL: https://api.sandbox.oscato.com/api/charges/26d5b179-a3b5-4aa3-ad98-3074b41e7d2b
• Basic Authentication: UBUY:BBtXa7yubR8ipuT
• Method: DELETE

A successful cancellation will return an HTTP 2xx return code. Otherwise it means it failed.

A cancellation can fail if, for example, the transaction was already locked by the provider because it is about to be processed. The best strategy in such a situation is to wait until the payment is processed, then issue a normal refund.

Note: This cancellation is similar to the cancellation of a Deferred Charge, but the underlying situation/status of the transaction is different.

Recurring Charges

In every integration scenario, you will get the customerRegistrationId and customerRegistrationPassword via Status Notification when you register a customer for the first time (by registering their first payment account).

The customerRegistrationPassword will only be sent ONCE by a notification with entity=customer by which it should be stored. But the customerRegistrationId is still sent with entity=payment.

If a customer has payment accounts registered for recurring charges, you can trigger those recurring charges by posting to:
/api/customers/<customerRegistrationId>/charge

You do not have to specify the channel, it is set to RECURRING by default for this type of call.

Provide the password via customer.registration.password. Registration ID is passed via URL path and is not required in this scenario.

URL: https://api.sandbox.oscato.com/api/customers/aff7a069-b77a-4023-b5ae-ad4f48596ce4/charge

• Basic Authentication: Merchant code and Token
• Method: POST

1. As mentioned above, there is an additional Boolean attribute hasRecurringAccount=true|false in the notification with entity=customer. If this customer has given the recurring mandate (via allowRecurrence=true in the CHARGE Request) and not revoked it (by removing the recurring account) hasRecurringAccount will show true; otherwise false.
   • Note: true does not guarantee the a recurring charge will work (or has ever worked), it just means that a mandate was given.
2. When there is a change in recurring charge status, a new notification will be sent.

• Example 1: If the customer had a registered "frontend" account, then registers a new account for recurring charges, there will be another notification.
• Example 2: If the customer already had a recurring account and registers another one, there will not be a notification (because status will remain true, ie no change)

3. While an account registration is pending, it will be not affect the recurring charges. The registration state can only change after the registration is successful. This may mean a notification for a newly registered customer (with regId and regPwd) will be sent. Soon afterwards, another notification with updated hasRecurringAccount info may also follow.

Automated Denial Management

• The Status Code will be pending and the reason code retry_scheduled.
• The Interaction Code and reason will be PROCEED - SCHEDULED, because on merchant side no action has to be taken. The returnCode.name will indicate why the transaction was unsuccessful to allow statistical analysis.
• The date and time of the next retry will be given in the notification, as a parameter named autoRetryAfter instead of retryAfter, but in the same format.
• declinesCount in the notification will state the number of rejections that have occurred for this payment request yet, starting with 1.

• The Status Code will be declined and the reason code debit_declined.
• The Interaction Code and reason will be the original codes that resulted from the final payment attempt.
• No autoRetryAfter or retryAfter value will be returned.
• declinesCount in the notification will state the number of rejections that have occurred for this payment request.

• The Status Code will be charged and the reason code debited, as with a normal successful Charge
• The Interaction Code and reason will be the original codes that resulted from this payment attempt, i.e. PROCEED and OK.
• No autoRetryAfter or retryAfter value will be returned.
• declinesCount in the notification will state the number of rejections that have occurred for this payment request until now (and therefore in total).

The precise retry schedules can be individually fine-tuned in their configuration depending on the Interaction Reason Code.

declinesCount parameter is provided to inform the merchant about the number of retry attempts made for the payment. Once this number reaches 10 (system default), the payment account is automatically blocked by the system. The logic of increasing and resetting the count is described in this chapter below. This logic applies not only for retry attempts made by ADM but for merchant initiated recurring charges as well.

Blocking an account after 10 attempts is default system behaviour, but there is a possibility to switch off this logic, so the account will not be blocked. In order to switch off retry attempts count, please contact our support team.

The defaults are:

In some cases it makes sense to stop the OPG from repeating charges, e.g., if payments have arrived through another channel. You can do so by sending an HTTP DELETE to the Charge resource (in OPG1: the retry entity associated with the Charge resource).

Example Request

• URL: https://api.sandbox.oscato.com/api/charges/efdfc6c7-a803-4f8b-8e30-ef6d9b3d6b6e
• Basic Authentication: UBUY/BBtXa7yubR8ipuT:
• Method: DELETE

As a reply you will get an HTTP 200 status code with an empty body. This indicates that the OPG will not repeat the scheduled charges for the referenced initial Charge any more.

This will also trigger a Status Notification with these values (among others):

• statusCode = canceled
• reasonCode = debit_canceled
• interactionCode = ABORT
• interactionReason = UNKNOWN

If you dig into the logic of ADM in the context of Fallback Routing (i.e., different providers that can be used for recurring charges), you will run into some conceptual challenges, because different providers could return different decline reasons.

Here is how OPG handles a situation like this under the hood:

Let's assume a recurring CHARGE for a registered customer that has 2 registered payment accounts (VISA and MASTERCARD). Let say that VISA has two PSP registrations for recurring charges, PROVIDER1 and PROVIDER2; and MASTERCARD only one via PROVIDER1.

In the first step, the OPG will try to process via VISA card, as in this example, this network has higher priority for some reason. If the processing via PROVIDER1 fails, the OPG will analyse the result. If the failure is 'recoverable' (Result Codes like 'provider not available' or 'provider internal error') the OPG will try to process via PROVIDER2. If the error is not recoverable the OPG will skip the processing via the other route (other provider registration) of the current registered account (in this case VISA) and will try to process with another registered account (in this case MASTERCARD). The OPG will also analyse the error result code to apply this information to the denial state of the registered account. 

As an example, let us say that processing with VISA card was tried and failed for both PSP registrations. For PROVIDER1 with Result Code 'provider not available' (recoverable) and for PROVIDER2 with 'exceeds limit'  (not recoverable). OPG will calculate an interaction code for the last try (here: PROVIDER2) and use it to increase the payment account's denials counter by 1 and change the account's denial state. This will also block the payment account if its denials counter reaches its limit (system default: 10). If the payment was successful the OPG would reset the account's denial counter to 0.

After that the OPG will try to process the transaction with the other registered account (MASTERCARD). Let say this try also fails with the Result Code 'provider not available' . As before, the OPG will derive the interaction code, use it for denial management, and return it to the merchant.

Now the transaction processing failed for both registered accounts VISA with Result Code 'exceeds limit'  (second result code, coming from PROVIDER2) and MASTERCARD with 'provider not available'. The last try will be taken as final execution result. Based on this result the OPG will derive the final interaction code and use the ADM to find the suggested next retry time. As a result ADM could return: 'in 2 hours'. The OPG will increase the rejection counter for this payment process (which is separate from the payment account mentioned before) and will schedule the retry event, via the scheduler component, in 2 hours.

After 2 hours the scheduler will fire a retry-event and we will try to process this CHARGE again. If this try also fails, the OPG will retry with the same logic again until ADM blocks the next retry and finishes this CHARGE process with actual denial handler's 'failure result interaction code'.

Deferred Payments

The reservation has a different commitment and duration depending on the payment network and/or the issuing bank. For credit cards, a reservation ("preauthorization") typically lasts 30 days whereas a successful closing of this amount is guaranteed for 10 days. For PayPal the reservation lasts 29 days although for a direct debit a successful closing is not guaranteed.

The same process is used for invoice payments. Typically the closing is called as soon as the products are shipped out to the customer. In retail businesses this can take days or even weeks. For Open Invoice Providers this is the signal that the payment is due, whereupon they should transfer it to you and start waiting for the invoice payment of the end customer.

The diagram below shows the deferred payment based on the AJAX Integration scenario:

Initialize a Deferred Payment

If a payment network supports deferrals then it needs to be configured in case you want the deferred payment to be enabled. The configuration would be done by our staff individually per merchant (within the limitations of that network).

You can enable the payment networks that are in "deferred", "non-deferred" or "any" mode, as explained below in the LIST Request, by adding the preselection.deferral parameter with one of the following values:

• DEFERRED Only networks which support deferred payments are listed. They are set in deferred mode for subsequent charges
• NON_DEFERRED (default) All networks which support immediate payment collection are listed. They are set in non-deferred mode when a CHARGE Request hits them
• ANY All networks are returned. If they support both modes, they are returned in deferred mode.

Note: The system default is set to "non-deferred"; therefore, if a payment network is configured to be in "deferred" mode and the value is not specified in the preselection.deferral parameter in the LIST Request, then the LIST Response would not include this payment network.

As with a normal LIST Request, you will get back all available networks together with their operation URL, which in Native Integration you interpret according to LIST Response.

It is possible to see the status code in the PURE_NATIVE integration scenario only, not with the other integration scenarios for security reasons. You could use the PURE_NATIVE integration parameter for testing reasons, but normally you would use the integration scenario that applies to your case. 

Make the Reservation

Now you can submit the customer's account details to the operation URL of the selected network. If the network is in deferred mode, this results in the "reservation" of the amount and a DEFERRED CHARGE object will be created. The Status Notification for this will return the statusCode = preauthorized.

This DEFERRED CHARGE could either be canceled or closed in a second step, see below. Both actions are triggered from server-side.

For security, only in "Pure Native" integration scenarios will your backend receive the required closing URL directly in the LIST reply as shown on the right side. In other cases it can be constructed, see below.

Cancel the Reservation

Whenever your system realizes that a payment is no longer required, because, for example, you cannot deliver the goods, it should cancel the (remaining) reservation. Otherwise your customer may not be able to use the blocked funds on their account until they expire (ie after 10 days). For some payment networks (Open Invoice methods, for example) these requested funds may never expire. Future payments may therefore lead to rejections if high amounts are left open.

To cancel a reservation, issue an HTTP DELETE request on the DEFERRED CHARGE resource. In "Pure Native" scenarios you get the URL to "self" already with the CHARGE reply.

Generally it is constructed like this:

https://api.sandbox.oscato.com/api/charges/[longId of DEFERRED CHARGE]

As a result, the reservation will be canceled through the provider and the transaction will transition to the status code canceled / preauthorization_canceled. This will be reflected in the response, which returns the DEFERRED CHARGE object again. Note that this DEFERRED CHARGE object is not actually deleted, you can fetch its status any time and it remains visible in My Transactions in the portal, for example.

It is possible to close a payment partially and cancel the remaining amount. To do that, you would have to delete/cancel the DEFERRED CHARGE after the partial closing has been made. This would automatically cancel the rest of the amount.

This order should be in line with most merchants' workflows, especially since cancellations of preauthorizations are not time-critical.

Close the Transaction

To complete (capture) the transaction and actually trigger the payment, you issue a POST request from your backend to the closing URL. It can be constructed like this:

https://api.sandbox.oscato.com/api/charges/[longId of DEFERRED CHARGE]/closing

In simple cases, the request contains only an empty  JSON object {} which means that the closing will be performed on the full amount.

But the closing can also be performed partially and multiple times. You can do this if, for example, an order has to be split into different shipments. In this case, you can submit a different amount and also a list of products covered by this payment.

A successful CLOSING operation will create a CHARGE object on the Orchestration Platform side, which represents the actual capture of the funds. It has a different ID from the original DEFERRED CHARGE object, and your system will receive a response and a notification about its status. If your system issues multiple partial closings, the resulting CHARGE objects will all have different IDs.

A CHARGE object in the response directly includes its own PAYOUT link. Your system should issue PAYOUTS on the ID(s) of the corresponding resulting CHARGE(s), not on the preceding DEFERRED CHARGE object.

For invoice payments, a partial CLOSING is the signal to expect less money transferred, eg due to a return of goods by the customer.

Chargebacks

Chargebacks are by definition asynchronous. Therefore, the OPG will send a Status Notifications to inform your system about them. This means simplified implementation for you, because every payment provider delivers this information differently, and often only passively, meaning active polling is needed.

Before an actual chargeback takes place, a payment provider can also request clarifications from the merchant, or announce a scheduled chargeback while still giving the merchant a chance to intervene.

These stages are modeled and described under Status and Reason Codes. Below is an representation of their possible transitions:

Fallback Routing

In general, certain Return Codes from providers are considered recoverable errors, for example, technical failures or configuration problems. Others, like stolen credit card, exist independently of the provider and are not considered recoverable in this context. See Return Codes for a full list.

If a recoverable error is detected, the payment gateway will look for an alternative provider contract that is also configured to be used to process this transaction. This means it needs to support the same parameters, like selected payment network and given country.

For frontend charges with customer interaction, fallback routing is straightforward. If the provider declines the transaction for a recoverable reason, the payment data as provided by the customer will be used to immediately try again with another provider. If one attempt succeeds or all available routes fail, the CHARGE Response and Status Notification with the result from the last attempt will be sent to the merchant.

Charges without customer interaction, ie Recurring Charges, are handled differently because the Payment Gateway is not allowed to store credit card verification numbers (CVC and CVV) and cannot therefore simply turn to a new provider as a fallback. From a providers' perspective, it would be an initial payment attempt requiring the verification number, which cannot be provided.

Therefore the payment gateway depends on past registrations by that customer, with alternative providers. If those are present, they will be used.

If no route through available registrations succeeds, the payment gateway will fall back to alternative registered payment accounts, another credit card that is stored, for example. If those are found, the whole story will be repeated with these accounts, ie looking for alternative providers again.

Since a direct reply with error information is required instantaneously for fallback routing, the logic does not work after an asynchronous provider or payment network is contacted. 'Asynchronous' in this context means only a temporal pending status is returned to the payment gateway, the final result arrives later. This is the case for providers like Adyen or Dotpay, or for redirection networks where the customer is sent to an external page.

But if no technical connection to the asynchronous provider could be established in the first place, owing to a provider downtime for example, Fallback Routing will be triggered because the OPG is able to act instantly.

Gambling Restrictions

References & Resources

ID and Object Types

Transaction ID Types

A transaction object contains these IDs:

• Assigned by the Orchestration Platform (will alway be present):
  • longId: generated ID (uniqueness guaranteed) for consecutive operations
  • shortId: generated string that's rather short and "almost unique" (i.e. not guaranteed for all times) that can be used for simplified human to human communication, e.g., over the phone
• Assigned by your system:
  • transactionId (mandatory for LIST requests) The primary, merchant-assigned ID (no uniqueness required or checked by the Orchestration Platform)
  • invoiceId (optional) An additional merchant assigned ID, mostly used by invoice based payment methods
• Assigned by payment providers (will not always be present):
  • pspId An ID assigned by the payment provider (if the transaction reached the provider)
  • institutionId An ID assigned by the banking institution in the background (if the transaction reached the institution and the ID is revealed by the PSP in front of it)

You will find these attributes inside the identification section of a transaction object (for authenticated server-to-server requests), except invoiceId which is inside the payment section.

Transaction Object Types

Note that the operations PRESET and cancel do not create separate transaction objects, that's why there are no corresponding objects listed here. They are only modifications on existing resources.

Other ID and Object Types

Apart from the resources that represent payment transactions there are also other object types and IDs, most notably for customer and payment account registrations:

• CUSTOMER objects: Are created by CHARGEs with account registration. They are identified by the customerRegistrationId. They can be used to issue Recurring Charges.
• ACCOUNT objects: A registered customer can have multiple registered payment accounts, which are identified by their accountId. They will be used implicitly for payments with registered accounts or updates or deletions of accounts. They can be also used explicitly to set preferred accounts. Or, after a LIST with "channel": "RECURRING" there can be Recurring Charges issued directly to a registered account.

Status Notifications

Technically the notifications are normal HTTP(S) GET requests with appended query parameters (as if they would come from a browser). They are sent to a URL defined by you, either by configuration or passing a notificationUrl parameter in the LIST Request. In other words, the OPG acts like it would request a web page from your server system and delivers the status parameters this way.

When a payment operation is triggered, there will be both, a synchronous reply to that request, and an asynchronous notification. In AJAX, Hosted and Display Native integrations the reply will arrive at the customer's client (e.g., browser), whereas the notification will directly arrive at your servers. Therefore it is safe to trust the notifications, as they never go through the customer's client.

The client should simply react on the reply, displaying the result to the customer. Your backend should trust the notification (only in Pure Native scenarios this difference is obsolete, because the reply also goes to the backend and can be used instead of the first notification).

There is one exception though: If an asynchronous PSPs is used (e.g., Adyen), the Interaction Code in the reply could be PENDING (and also a notification with the Status Code  pending will arrive). In this case the customer should be asked to wait. Typically after a few seconds (in rare cases maybe even half a minute) a new notification will arrive at your server with an updated Status Code. Only for these cases there should be some kind of synchronization with the customer's client.

It is important to listen to status notifications, also after a payment is completed, using the Status Codes to note events like chargebacks.

Note that a notification can concern different things as indicated by the value of the entity attribute:

• payment: The payment transaction (relevant for successes, chargebacks, etc.)
• customer: The customer registration (as a consequence of account registration)
• account: The customer's account registration (network name and account registration ID, as a consequence of account registration)
• session: The frontend payment session (relevant when deciding whether to issue a new LIST Request or update an old one)

In the simplest implementation you will only need the payment notifications. Therefore do not confuse the entities.

Note: not all parameters will always be present since status notifications could be pushed in different cases. For example, in case an account is registered, a status notification would be pushed without the "amount" parameter since it is independent from the amount. Also, the order of parameters is not fixed.

mandate data: Was introduced for SEPA network and returns information about the official mandate data that was used for a transaction.

remittanceAccount[n] data: Provide this to your customers so they can complete their payment in case of invoice or prepaid methods. Only the fields required for the payment will be provided. For example a bank account together with a reference text (aka descriptor) could be identified, where the customer should transfer their money to. There could be more than one possible remittance account (indicated by [n] below, starting with remittanceAccount[0]).

If you have not implemented the notification listener already, do not worry about these values, which are either deprecated or not fully in place, yet.

The delivery of notifications is guaranteed (for any interface version). This means:

1. A notification delivery is considered successful if the target host (merchant server system) replies with an HTTP 2xx status code within a specified timeout allowance, currently 10 seconds (but may be subject to change).
2. If a notification fails, it is repeated in progressively increasing time intervals until successful. The first repetition happens after 2 minutes, then 5, 10, 15, and further.
3. The last delivery attempt takes place around 24 hours after the first try. If this also fails the error is considered permanent and the Payment Gateway stops repeating this notification.

Please consider the following implications of guaranteed delivery:

1. If a notification is correctly received by a merchant server system, but the response takes longer than the OPG's timeout, the message would be regarded as failed and repeated later. Therefore your backend will receive the same notification twice, i.e., a duplicate. Duplicates can be detected using the notificationId, which will stay the same across all repetitions.
2. If a notification (e.g., conveying a PENDING Interaction Reason) fails, and a few seconds later a new notification (e.g., with OK Interaction Reason) is successfully delivered, the first one will be re-sent 2 minutes later. In the given example this means the PENDING notification would arrive after the OK message. This information, of course, is not correct anymore. Therefore the notification order should be checked based on their timestamp. For example, the timestamp of the last notification could be saved and only notifications with higher timestamp will be processed in the future.

Note: These examples are unlikely, especially example 2, but they could occur. Therefore we recommend a clean implementation regarding these cases in the long run.

Also, if some kinds of notifications are ignored by your system (e.g., with entity = session, because you do not need payment conversion tracking), please let the system still return an HTTP 200 OK, so that the guaranteed delivery does not take effect.

Keeping your notifications secure

It's important that you can recognize the notifications as valid in cases when a malicious party tries to spoof your servers by generating false notifications. You can do it in one of the three available ways:

• by adding custom parameters to the notificationURL (shared secrets)
• by opting for POST notifications
• By adding additional headers to your notifications

If, for notification security reasons, you allow-list our IP's, note that we stop supporting static IP's as of April 1, 2021. Use one of the methods mentioned above to make sure your notifications are secure.

Custom Parameters (shared secrets)

You can  add custom name=value pairs to the list of parameters in the notification URL by passing them in the LIST request callback.notificationUrl definition. Any number of parameters added to the notification URL in the LIST request will be sent back once the notification arrives in your servers. On the right pane we can see the example of a token containing the value 512938z1b176598113bA1b3 sent through a LIST and later received as a notification parameter.

Use the shared secret method to control if the notifications sent by the OPG are expected by your systems. You can achieve it by adding a token to URLs provided to the Orchestration Platform and checking that notifications received through these URLs have the same token. 

Additional headers

You can add more metadata to your transactions by providing custom headers to the transaction LIST request. The headers will be returned as a list in the notification you receive from our orchestration platform. The headers can be added on the transaction or division level.

Note that unlike the header "value",  the header "name" cannot contain any special characters (such as spaces or underscores). 

You can use this additional information to separate divisions in your systems, or for analytical or security purposes.

There is no limitation to the number of headers other than that of your HTTPS server.

There are also few limitations on the content of the header and we accept everything that doesn't overwrite OPG original headers. For example, we don't allow user agent number to be in the header. If there is a risk that the original header will be overwritten, the additional header is dropped by the system.

POST and GET notifications

You can choose to receive notifications as a GET (the default) or POST request. In GET, the data will be placed in the query URL and in POST, in the body. Fore security reasons, we advise selecting POST notifications.

HTTP Request Inspector

Interaction, Status, & Return Codes

After an operation, the OPG provides different information in form of codes or messages that can be used by the merchant systems to understand what might be the next operation step and what is the current state of an operation. These codes and messages should not be presented to end customers due to their technical content.

There are 4 different code types sent during a payment session: Interaction codes, status codes, return codes, and result codes, plus a longer message delivered through the result info field.

The matrix below summarizes them:

In the next sections you can find detailed information about all codes described above, and in which situations they should be used.

Interaction Codes

The interaction code is actually composed of two parts: the code itself and a reason, which expands its meaning and can be used for more refined decision.

Interaction Codes are designed to lead a merchant shop system to the appropriate action in a payment session. However, they are not a comprehensive reflection of all the statuses a payment may have. This information is present in the Status Codes, which bring a deeper understanding of the current status of a transaction. Keep in mind that the Status Codes don't need to be evaluated during a session, but they can be helpful case a merchant wants to investigate the details of a particular scenario while troubleshooting or testing. 

Status Codes

The whole Status is composed of a Status Code and a Reason Code, which provides further information inside a group of Codes.

Note that a status can concern different scopes as indicated by the value of the entity attribute:

• payment: The payment transaction (relevant for successes, chargebacks, etc.)
• customer: The customer registration (as a consequence of account registration)
• account: The account registration under a customer (also result of account registration)
• session: The frontend payment session (relevant when deciding whether to issue a new LIST Request or update an old one)

Status Codes are complemented by Interaction Codes that offer recommendations on what to do next in a frontend payment session, especially to reduce declines in problematic situations.

Conceptually the status codes can be seen in a hierarchy as visualized below:

Please note that the diagram above, while being a general classification of payment statuses, does not represent the actual transition flow expected in a payment session. In fact, many PSPs can initially reject a payment request (due to internal, identified errors) and later accept it. Also, initially approved transactions can move in a later stage to a rejected state (example chargebacks). For this reason, the OPG will always listen to the PSPs' backend notifications and update the transaction statuses accordingly, even transitions between such "positive" and "negative" states.

Concerning entity = payment

Concerning entity = payment

The following statuses can occur for Deferred Charges or related Closings, in addition to the statuses above.

Concerning entity = payment

The following statuses can occur for PRESET flows in Delayed Payment Submission and Express Checkout, in addition to the statuses above. Note that in Delayed Payment Submission the preset account is part of the LIST object which will not change the status. Therefore these statuses only occur on the PRESET call itself in this case, and are visible if issued from the backend.

Concerning entity = customer

The customerRegistrationPassword will be sent exactly once after an account registration with such a notification. customerRegistrationId will be returned for other notifications as well.

Concerning entity = account

Concerning entity = session

Result Codes

Result Codes describe the outcome of an operation in detail. To simplify, the OPG performs a mapping of the Return Code to interpret the Interaction Codes, which we recommend to use in most cases.

A Result Code consists of three groups separated by dots:

1. The first group is numeric and called Return Code (or formerly Standard Code). The Return Code 00000 means that no error occured during payment request processing.
2. The second group contains alphanumerical characters and refers to a component that generated this return code. Standard components are described by three digits whereas custom components and server extensions are usually referred to by their code names.
3. The third group is a qualifier that may contain additional information, as given by the provider. Usually the qualifier is encoded by three digits and default value is 000 if no additional information is available.

Component and qualifier are called Investigation Info and help to analyze error conditions. Result codes are complemented by Result Info messages, which contain original payment service provider information, when available.

Result Code: 00100.102.000
Return Code = 00100 - Invalid request
Component code = 102 - Merchant HTTP Gateway Listener
Qualifier = 000 - no additional information

Return Codes

• returnCode: numerical representation, shown as part of the Result Code.
• returnCodeName: human readable representation, see table below.
• returnCodeSource: indicates where in the chain the response was generated. Together with the pspCode and institutionCode this should make clear, which organization should be contacted in case of transaction specific questions. Its value can be either of:
  • GATEWAY (Payment Gateway, first level)
  • PSP (Connected payment service provider, second level)
  • INSTITUTION (The banking institution that may be behind the payment service provider)

The table below includes all Return Codes of the Payment Gateway with their numbers, names, descriptions and if they are considered recoverable in the context of Fallback Routing.

Test Accounts & Magic Numbers

Test Accounts

Please note that we don't keep track of updates of PSP test card numbers, so the table below might not be entirely correct. To avoid false negatives while testing, please make sure that you get the PSPs' Sandbox test numbers directly from their developer documentation. We link some of these PSPs docs directly under their name in the table below.