LLM Documentation Integration

About 583 wordsAbout 2 min

2025-03-27

PingPong Checkout documentation already exposes structured AI-facing entry points. You can use llms.txt / the llms-full.txt compatibility entry / OpenAPI for document discovery, and optionally install the Skill when you want the AI tool to follow PingPong Checkout-specific workflows and checks.

Which option should you choose first

Goal	Recommended approach
Let AI read the latest docs, API descriptions, and page links	Start with `llms.txt` + `/api/openapi.json`
Support tools that still read `llms-full.txt`	Use `llms-full.txt`, which returns language-specific `llms.txt` discovery entries
Let AI work with PingPong Checkout-specific flow, review criteria, and best practices	Install `pingpong-checkout.skill`

Discovery endpoints

File	Path	Purpose
llms.txt	`/llms.txt`	Root entry that returns the language-specific discovery files
Chinese llms.txt	`/zh/llms.txt`	Chinese documentation index for page-by-page retrieval
English llms.txt	`/en/llms.txt`	English documentation index
llms-full.txt	`/llms-full.txt`	Compatibility entry that returns the language-specific discovery files
OpenAPI Spec	`/api/openapi.json`	OpenAPI 3.1.0 specification for schema and parameter validation
API Index	`/api/index.json`	Exported API catalog with links back to JSON and Markdown docs

Recommended retrieval order

Start with the locale-specific llms.txt
Read /api/openapi.json for endpoint constraints and schemas
If your tool is fixed to llms-full.txt, follow the language-specific llms.txt links returned there
Continue with the page links exposed by llms.txt or api/index.json

AI tool integration examples

Cursor

Create a .cursorrules file in your project root:

When calling PingPong Checkout APIs, prefer the following references:
- English Documentation Index: https://acquirer-api-docs-v4-en.pingpongx.com/en/llms.txt
- OpenAPI Specification: https://acquirer-api-docs-v4-en.pingpongx.com/api/openapi.json

Windsurf

Add to .windsurfrules:

PingPong Checkout API Documentation:
- https://acquirer-api-docs-v4-en.pingpongx.com/en/llms.txt
- https://acquirer-api-docs-v4-en.pingpongx.com/api/openapi.json

Claude Code

Configure in CLAUDE.md:

## PingPong Checkout Documentation Reference

When consulting PingPong Checkout APIs:
1. Prefer /en/llms.txt for the English documentation index
2. Read /api/openapi.json for schemas and parameter constraints
3. Follow the linked documentation pages for detailed guidance

Generic HTTP access

Any AI tool that can make HTTP requests can fetch:

# Get the language index
curl https://acquirer-api-docs-v4-en.pingpongx.com/llms.txt

# Get the English documentation index
curl https://acquirer-api-docs-v4-en.pingpongx.com/en/llms.txt

# Get the llms-full compatibility entry
curl https://acquirer-api-docs-v4-en.pingpongx.com/llms-full.txt

# Get the OpenAPI specification
curl https://acquirer-api-docs-v4-en.pingpongx.com/api/openapi.json

# Get the API index
curl https://acquirer-api-docs-v4-en.pingpongx.com/api/index.json

What you get from the API index

/api/index.json returns the currently exported API catalog. Common fields include:

Field	Description
`version`	Index format version
`generated`	Index generation timestamp
`count`	Current number of exported APIs, which changes as docs evolve
`apis[].path`	Endpoint path such as `/v4/payment/prePay`
`apis[].method`	HTTP method
`apis[].jsonUrl`	Structured JSON description
`apis[].mdUrl`	Markdown source URL
`apis[].permalink`	Site page URL

Skill vs. document discovery

llms.txt / llms-full.txt / OpenAPI answer the question “how does the AI find the documentation?”
pingpong-checkout.skill answers the question “how should the AI work through a PingPong Checkout task?”

Use discovery endpoints when you want the AI to find APIs, parameters, and documentation pages. Add the Skill when you also want guided review, workflow discipline, risk reminders, and pre-launch checklists.