---
title: "How it works"
date: 2026-07-24
author: "Claude MCP"
---

# How it works

Once enabled, every published article and category on your site is also available as plain Markdown, for AI agents, LLMs, and other tools that prefer plain text over HTML.

 
## Requesting the Markdown version

 A visitor, or a bot, can get the Markdown version of any article or category page in three ways:

 
- Add **.md** to the end of the page's URL, e.g. `example.com/my-article.md`.
- Add **?output=markdown** to the URL, e.g. `example.com/my-article?output=markdown`.
- Request the page with an **Accept: text/markdown** header.

 This only applies to article and category pages. Other page types aren't affected.

 
## What's in the response

 The response starts with a short frontmatter block (title, and any of date, description, author, images, category or tags you've enabled in **Configuration**), followed by the content converted to Markdown. Headings, links, images, lists, tables and code blocks are all converted, and any Joomla plugin shortcodes are stripped out. If **Show Custom Fields** is on, filled-in custom fields are added as a "Custom Fields" section at the end.

 The response also includes an **X-Markdown-Tokens** header, a rough estimate of the token count, useful for judging how large a page is for an AI agent before fetching it.

 
## Visibility and access

 The Markdown version follows the same rules as the normal page: unpublished articles, articles outside their publishing dates, and content the current visitor's access level doesn't allow are never included.

 
## Troubleshooting a missing field

 If a custom field isn't showing up in the Markdown output, request the page with **&debug=1** added to the `?output=markdown` URL, e.g. `example.com/my-article?output=markdown&debug=1`. This returns a plain-text report of what the plugin found for that article, including any custom field values it read from the database, useful for spotting a field that's empty, unpublished, or assigned to the wrong context.