3 ms·
Should every web page expose an AI-friendly JSON representation?
My website already includes AI-related files such as llms.txt.
I'm considering creating a separate JSON file for every page and article so AI systems can understand the content more easily and accurately. I would reference this JSON file from the page's <head> using a <link> tag.
The JSON file could contain information such as:
Page URL
Canonical URL
Title
Summary / Description
Main Content (clean article content)
Author
Published Date
Last Updated
Entities (people, companies, places, products, etc.)
Keywords / Topics
FAQ
...
My idea is that AI crawlers could read this structured JSON instead of having to extract the main content from noisy HTML that contains navigation menus, sidebars, ads, comments, JavaScript, tables, and other non-essential elements.
I have two questions:
Could this approach reduce the chances of AI crawlers misunderstanding a page or extracting incorrect information from HTML, advertisements, tables, comments, or other noisy content?
Do you think a page-level JSON file like this could help AI systems better understand a page and potentially improve AI recommendations, citations, or other AI-generated responses in the future? Why or why not?
- arbol 1mo agoThis is already happening and it absolutely does let llms understand the page better. Search for JSON-LD
- SeenSure 1mo ago[flagged]
- massimodeluisa 1mo agoJSON-LD already covers a chunk of that list inside the HTML, so a separate AI-friendly JSON file helps mostly when the page itself is too noisy to parse (don't tell me )... What still sits underneath is earlier than extraction:: can the crawler fetch the main HTML at all, and is there readable text once the JS shell is gone? "A twin behind a link tag" does not help if the UA never gets a readable document. BUT there's IsReady.AI that help you creating AI readable content of your website. Plus consider firecrawl.dev!
- benban 27d ago[flagged]
- SEOvaultai 24d ago[flagged]