Greetings programs,
There’s an emerging convention called
llms.txt,
a plain text manifest designed to expose structured documentation links for retrieval by large language models. It allows developers and organizations to explicitly declare which markdown, API references, or guides should be indexed by AI assistants such as
Context7 or GitHub Copilot’s contextual search.
Although adoption is still early,
llms.txt offers
a clear advantage for documentation platforms like
https://jakarta.ee/specifications/ : it provides a standard way to control what AI systems see and use,
ensuring that generated answers reference current, authoritative Specification text rather than scraped or outdated copies. This can reduce hallucinations and improve the quality of AI-driven support, tutorials, and code suggestions.
The canonical reference is here:
https://github.com/AnswerDotAI/llms-txt
Given Jakarta EE's scale and influence, evaluating support for
llms.txt—even
experimentally—would help shape best practices for AI documentation accessibility and discoverability.
I am not in a position to drive this forward, but I at least wanted to bring it to your attention.
Thanks,
Ed
_______________________________________________