Serve Markdown content on Accept: text/markdown requests for full AI Content score - Part 2

Following on from the llms.txt feature request — once llms.txt and llms-full.txt are in place, the remaining gap to a full score on the Content dimension of isitagentready.com is Markdown content negotiation.

When an AI agent requests a page, it sends an Accept: text/markdown header. If your server returns HTML, the agent has to parse out navigation, scripts, ads, and markup noise to get to the actual content — burning tokens and often getting it wrong. If your server detects that header and returns clean Markdown instead, the agent gets exactly what it needs, efficiently.

The request:

When eDirectory detects an Accept: text/markdown request header, return a clean Markdown version of the page instead of HTML — covering at minimum:

• Category pages (name, description, listing summaries)

• Listing detail pages (business name, description, contact info, category)

• Key static pages (About, Contact, Add a Listing)

This doesn't need to replace the HTML response for normal browsers — just intercept the header and serve an alternate clean representation to agents that ask for it.

Why it matters:

This is the difference between an AI agent being able to read your directory and being able to use it as a reliable source. Combined with llms.txt, a proper robots.txt with AI bot rules, and a sitemap, this completes the first three AI readiness categories and puts eDirectory sites at the top of the scoring curve.

Most platforms haven't implemented this yet. It's a genuine differentiator right now.

Aaron