As AI continues to reshape how we build and interact with software, the SAP Cloud Application Programming Model (CAP) is embracing this shift by introducing two new files: llms.txt and llms-full.txt. These files are designed to enhance the way large language models (LLMs) like SAP Build Code or ChatGPT or any AI IDE understand and interact with CAP documentation and projects.
In this blogpost, we’ll explore what these files are, how they benefit users, how to integrate them into your workflows, and what limitations you should be aware of.
LLMs.txt is a new web standard designed to help reasoning engines understand website content better. It addresses the limitations of AI systems by providing a format optimized for AI processing, removing non-essential information, and presenting content more efficiently. The concept was popularized by Mintlify, which added support for LLMs.txt to their docs platform. The file format includes two types: /llms.txt for navigation and structure, and /llms-full.txt for comprehensive documentation. Several tools are available to help generate these files, and their adoption is expected to grow as AI-readability becomes crucial for web content.
LLMs.txt is a markdown file with a specific structure. The specification defines two distinct files:
Prompt Augmentation: When using an LLM, include a reference to https://cap.cloud.sap/llms.txt in your prompt. For example:
“Using the documentation at https://cap.cloud.sap/llms.txt, explain how to define a service in CAP.”
RAG Pipelines: If you're building your own AI assistant or chatbot, you can ingest the URLs from llms-full.txt into your vector database or document retriever to enable semantic search over CAP docs.
IDE Integration: Advanced setups can use these files to power AI-based code completion or documentation lookup tools within IDEs like VS Code.
While llms.txt and llms-full.txt are powerful tools, they come with a few caveats:
The introduction of llms.txt and llms-full.txt is a forward-thinking move by the CAP team, making it easier for developers to harness the power of AI in their workflows. Whether you're building intelligent assistants, enhancing developer productivity, or simply exploring CAP, these files offer a structured way to bring AI and documentation closer together.
Samples of /llms
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
User | Count |
---|---|
24 | |
12 | |
11 | |
10 | |
9 | |
7 | |
7 | |
7 | |
7 | |
6 |