Skip to main content

LLM (Large Language Model)

EasyPay Online Payment provides LLM-optimized training data and configuration guides to help AI-powered coding tools (ChatGPT, Claude, etc.) generate payment integration code more accurately.

Choose the method that best fits your development environment and complete complex payment integrations 10x faster.

Download AI Learning Resources​

Use the following files according to the tool you are using.

FilenameDescriptionURL
llms.txtContains only key API lists and documentation links.https://docs.kicc.co.kr/online/llms.txt
llms-full.txtFull text containing detailed specifications and examples for all Online Payment APIs.https://docs.kicc.co.kr/online/llms-full.txt
💡 What is llms.txt?

llms.txt is a standard Markdown-based file format designed to help AI models, rather than humans, efficiently learn the content of a website.

How to Integrate Online Payments using LLMs​

We recommend providing the URL to AI models with internet browsing capabilities, and the Full Text file to those without.

Scenario A: GPT-4o / Perplexity (Web Access Available)​

Including the llms.txt link in your prompt allows the AI to answer based on the most up-to-date information.

Prompt Example: "Write a payment approval logic in Node.js referring to the EasyPay specification below. Doc URL: https://docs.kicc.co.kr/online/llms.txt

[Requirements]

  1. Use the Axios library.
  2. Set timeout to 30 seconds.
  3. Log errors when they occur (Must mask sensitive data like card info)."

Scenario B: Claude 3.5 Sonnet / Local LLM (File Upload)​

File uploads often yield higher accuracy than browsing. Download the llms-full.txt file and upload it to the chat window.

Prompt Example: "(After uploading the file) This document is the full API specification for EasyPay. Based on this, create a Payment Controller using Java Spring Boot. You must include logic to automatically call the Network Cancellation API (Net Cancel) in case of approval failure (especially on timeouts)."

Essential Security Checklist​

Even if AI writes the code, the final security responsibility lies with the developer. You must verify the following items before deployment.

  1. Merchant ID (Mall ID) Hardcoding:
    • ❌ const mallId = "055...";
    • ✅ const mallId = process.env.EASYPAY_MALL_ID;
  2. Logging Sensitive Information:
    • Are customer payment details or PII (Personally Identifiable Information) masked before being recorded in server logs?
  3. Exception Handling:
    • Ensure that the payment status does not remain 'Indeterminate' or 'Pending' when a network timeout occurs by verifying it with the Transaction Status Check API.