LLM context files

SDK4j provides an llms-full.txt file following the llms.txt convention. This file consolidates the complete SDK documentation — all service areas, method signatures, parameters, return types, and code examples — into a single plain-text file optimised for LLM consumption.

The purpose of this file is to give any LLM (Claude, ChatGPT, Gemini, Copilot, etc.) full knowledge of the sdk4j API so it can assist you in writing integration code without hallucinating method names or incorrect signatures.

You can download the file here:

llms-full.txt

How to use it

The recommended workflow is to place llms-full.txt in your project folder and reference it directly in your prompt using the @file syntax supported by most AI-assisted IDEs (Cursor, Windsurf, VS Code + Copilot, etc.) or by attaching it directly to your conversation in Claude, ChatGPT, or any other chat-based LLM.

Example

Follow these steps to bootstrap a new SDK4J project using an LLM:

1. Create a new folder for your project, for example sdk4j-test.
2. Copy llms-full.txt into that folder.
3. Open your AI assistant (Claude Code, Cursor, Windsurf, ChatGPT, etc.) in that folder.
4. Use a prompt like the one below.

Sample prompt

Create a Maven project:
- Call it sdk4j-test.
- You must use sdk4j version 5.4. In @llms-full.txt you have the full documentation for this library.
- The goal is to create an application that implements the following logic:
  - Verify that the folder /okm:root/test exists.
  - If the folder /okm:root/test does not exist, create it.
  - Upload a document to that folder (the path of the document to upload should be configurable from the command line).
- Create a README.md file in Markdown format explaining how to run the application from the terminal.

The LLM will read llms-full.txt to understand the SDK4J API and generate correct, working code.