GMO Payment Gateway Releases "payment Development Documentation" for the First LLM Search and AI Search Support for PSP in Japan
~Supporting the efficiency and accuracy of system introduction and development through AI-based specification confirmation~

GMO Internet Group's GMO Payment Gateway, Inc. (President & Chief Executive Officer: Issei Ainoura, hereinafter referred to as GMO-PG) today released the "payment Development Document" with search and AI search functions using large language models (LLMs) in Online Payment Service" PG Multi-Payment Service" from March 30, 2026 (Monday).
This document was first released to some merchants and developers after its announcement on February 3, 2026, and was made after repeated improvements based on feedback from actual users.
This document will help existing merchants and developers improve development efficiency and accuracy when changing specifications. Companies that newly introduce "PG Multi-Payment Service" can grasp specifications and connection methods from the consideration stage, as well as confirm expected implementation patterns in advance, which is expected to speed up decision-making and reduce the development burden of design and implementation after implementation.

【Background and Summary】

As the use of generative AI expands, the use of LLMs for information search is also progressing in development sites. In addition to referencing documents on a page-by-page basis as in the past, there are an increasing number of cases where relevant information is checked cross-sectionally while using AI as an auxiliary tool. In these explorations, developers input questions into the AI, which then references the content of the document to get answers. Therefore, it is important that documents are structured in a way that is easy to understand not only for humans but also for AI.
On the other hand, Japanese online payment has a variety of payment method such as CVS Payment and QR code payment in addition to Credit card payment, which makes the payment specification more complex. Developers need to cross-check multiple conditions and assumptions as they design mechanisms to incorporate payment functionality into their systems, and misreadings or missed references in the specification can affect quality and development timelines.
In light of these changes in the environment, on February 3, 2026, GMO-PG announced a policy to redesign the document structure based on LLM search and introduce an AI search function ^(*1), which was released in advance to select merchants and development companies. This is the first (* ²⁾ PSP (payment service provider) in Japan to provide development documentation that combines structural design based on the premise of web search using LLMs and the introduction of AI search functions. We have taken into account the feedback from developers who used it during the pre-release period and made repeated improvements to this release.
Developers at existing merchants can achieve more efficient and accurate development when changing specifications or adding new payment method. Companies considering online payment can accurately understand specifications, connectivity methods, and constraints from the start of the study phase, as well as preview expected implementation patterns. As a result, decisions can be made based on clarifying estimates and design prerequisites, which leads to reduced rework and reduced development burden in the design and implementation phases after implementation.
GMO-PG will promote the advancement of Japan's payment development environment by providing the actual operational knowledge cultivated through payment processing of more than 22 trillion yen ^(*3) per year in a form suitable for the AI era.

• *1 Announced on February 3, 2026, "GMO Payment Gateway, the first PSP in Japan, has developed payment development documentation with LLM search and AI search" URL: https://www.gmo-pg.com/news/press/gmo-paymentgateway/2026/0203.html
• *2 PSP (payment Service Provider) that mainly serves the Japanese market, and is the first case to combine the redesign of the document structure based on the premise of web search by LLM and the introduction of an AI search function. According to GMO-PG.
• *3 Consolidated figures as of the end of December 2025.

[Comments from early adopters]

note inc. Taketo Yamada, Project Manager
This time, I saw the new manual from the developer's point of view. With the addition of the search function and AI assistant, we feel that it is easier to get to the required specifications and parameters, and to efficiently solve any doubts during implementation.
In addition, the configuration that can be drilled down for each use case and payment method is easy to use, and you can check everything from what you want to do to how to implement it without hesitation, so I think it is easy for those who are introducing "PG Multi-Payment Service" for the first time to grasp the overall picture.
The addition of language-specific samples and request response examples to the API reference, as well as the ease of viewing the transition diagram of transaction states, is a very welcome improvement that directly leads to a reduction in development effort.
The fact that the addition of "llms.txt" makes it possible to refer to AI is also an initiative that is in line with the trend of the times, and we look forward to its future development.

GMO MAKESHOP Co., Ltd. Takato Koshikawa, Manager of the Development Department of the MakeShop Business Division
Previously, implementing payment integrations required cross-reference and review specifications, which took a lot of time to research.
By consolidating it into the "PG Multi-Payment Service Developer Documentation", it is possible to complete everything from payment method selection to connection methods, API references, and use cases in one place, so it is expected to significantly reduce the number of man-hours for implementation.
As a result, the development of such a development environment is reflected in the speed of development of new features, and it will lead to the provision of rapid introduction of payment method in line with trends to shops using "makeshop by GMO".
It is also expected to contribute to shortening the recovery time in the event of a problem. We appreciate the fact that we can devote more resources to measures that directly lead to improved customer experience while keeping development costs down.
I feel that the environment has been created where developers can concentrate on the original design and implementation, and I look forward to further enhancements in the future.

[Features of this published document]

This document is structured around the following three points.

1. Structural design based on the premise of LLM search
In addition to expanding the use cases according to business models and business requirements, we have organized a collection of links to basic information and documents in the "PG Multi-Payment Service", and developed a "llms.txt" format for LLMs. This has created an environment where it is easy for AI to grasp the overall structure of the document.
In addition, each page can be referenced in "Markdown" format, and is provided as structured text that is easy for LLMs to read. We have also implemented a "Markdown" copy function to make it easier for developers to paste into AI.
These measures make it easier for AI to read the content after accessing the document, improving readability within the site. As a result, it is easier to make references based on specifications to developers' questions when searching for information using LLMs, which leads to improved answer accuracy.

Implementing 2.AI Search Functionality
We have implemented a search function that allows you to cross-reference, regardless of page structure or search keywords, and to refer to related information from the content and questions under consideration. In addition, we have a new "Ask AI" function that allows you to ask questions to AI optimized specifically for documents. This document is designed as a reference to generate answers based on specification understanding and help developers make decisions.^（※4）

• *4 The content entered in the "Ask AI" function may be shared with third parties involved in the operation and provision of the Service, stored for a certain period of time, and used for the purpose of providing services and improving quality. Please refrain from entering any personal or confidential information.

3. Features that enhance developer convenience
The API reference expands the sample code and provides functionality that can be executed in the browser. By creating an environment where you can check the operation on the spot, we have achieved a "try it out immediately" experience. The design allows the acquired code to be immediately transferred to the company's environment, contributing to improving the verification speed in the early stages of implementation.

In addition, the documents provided by independent websites for each purpose are integrated into a single website, making it easy to reach the information you need, and it is structured to proceed as a series of flows from consideration to design and implementation.

Based on feedback from early adopters, this document improves usability and ease of understanding, such as improving the UI of the API reference, expanding the sample code, improving search response speed, and supporting shortcut activation (Ctrl+K).

■For more information on the use of this document and specific usage methods, please refer to the owned media "PX+ by GMO" that supports the growth of businesses starting from payment. Please take a look.

OpenAPI-enabled payment development documentation released.
Introducing the PG Multi-Payment Service Developer Documentation
URL：https://www.gmo-pg.com/pxplus/journal/8233/

【PG Multi-Payment Service】

Online Payment Service supports more than 30 different payment method, including Credit card payment, CVS Payment, and QR code payment. Depending on the business type and growth phase of the business, we provide a flexible payment foundation. In addition, in addition to adopting a development method that is easy to design and implement based on OpenAPI specifications, it is equipped with high security and fraud detection functions to support online payment in a wide range of fields such as e-commerce, digital services, and BtoB transactions. At the same time, we are continuously developing documentation and reference environments to help developers get to the exact information they need.