Azey

AI agents from ChatGPT, Claude, Perplexity, and others read your content to generate answers — but they parse through HTML, JavaScript, and navigation just to reach it. Serve Markdown gives them a direct path: clean Markdown with structured metadata. When a crawler requests it — via an <code>Accept: text/markdown</code> header or a <code>.md</code> URL — your site responds instantly. No noise, no guesswork.
<h4>How It Works</h4>
Serve Markdown adds three capabilities to your WordPress site:
1. Content Negotiation 
When an AI crawler sends <code>Accept: text/markdown</code>, your site returns Markdown instead of HTML — no URL changes needed.
2. .md URL Suffix 
Append <code>.md</code> to any post or page URL (e.g. <code>example.com/my-article.md</code>) to get the Markdown version directly.
3. Markdown Auto-Discovery 
Every page includes a <code><link rel="alternate" type="text/markdown"></code> tag in the HTML head so crawlers can find your Markdown automatically.
<h4>What the Markdown Output Looks Like</h4>
Every Markdown response includes YAML frontmatter with structured metadata followed by your post content. Here is a real example:
<pre>
---
url: 'https://akumarjain.com/textexpander-year-in-review-2025/'
title: TextExpander Year in Review 2025
author:
 name: Ajay
 url: 'https://akumarjain.com/author/akumarjain/'
date: '2026-01-05T09:05:00+05:30'
modified: '2026-01-10T14:30:06+05:30'
type: post
categories:
 - 2025
 - Year in Review
tags:
 - '#ToolsIUse'
 - Productivity
 - Remote Work
image: 'https://akumarjain.com/wp-content/uploads/text-expander-year-in-review-2025.webp'
published: true
---

# TextExpander Year in Review 2025

In my daily work, I need to type the same words and phrases repeatedly...
</pre>
<h4>Features</h4>
Content Serving
<ul>
<li>Content negotiation via <code>Accept: text/markdown</code> header</li>
<li><code>.md</code> URL suffix on any post or page</li>
<li>Auto-discovery <code><link rel="alternate"></code> tag in page head</li>
<li>Each feature can be toggled independently</li>
<li>Choose which post types are exposed (posts, pages, custom post types)</li>
</ul>
Frontmatter and Metadata
<ul>
<li>YAML frontmatter with title, author, date, categories, tags, featured image, and more</li>
<li>Toggle individual metadata fields on or off</li>
<li>Add custom static key-value pairs (e.g., <code>license</code>, <code>language</code>)</li>
<li>Map WordPress custom fields (post meta) into the frontmatter</li>
</ul>
Access Control
<ul>
<li>Per-post opt-out via a sidebar checkbox in the editor</li>
<li>Exclude entire categories from Markdown serving</li>
<li>Exclude entire tags from Markdown serving</li>
<li>Only published posts are served — drafts and private posts are never exposed</li>
</ul>
Crawler Insights
<ul>
<li>Request log showing every Markdown request with timestamp, URL, bot name, and method</li>
<li>Automatic bot detection: ClaudeBot, GPTBot, ChatGPT, OAI-SearchBot, PerplexityBot, Googlebot, Bingbot, Cohere, Meta AI, Bytespider, Applebot, and more</li>
<li>Stats dashboard: total requests, requests today, unique bots</li>
<li>Filter log by individual bot</li>
<li>Configurable retention period with automatic pruning</li>
<li>One-click log clearing</li>
</ul>
Editor Integration
<ul>
<li>“Preview Markdown” button in the post editor sidebar</li>
<li>“View Markdown” link in the Posts and Pages list table</li>
<li>Meta box shows on all enabled post types</li>
</ul>
<h4>Performance</h4>
Markdown conversion only runs when specifically requested — regular visitors see the same HTML pages they always have. The crawler log uses a lightweight custom table with three safeguards against unbounded growth: time-based retention, a row count cap, and a table size cap.
<h4>Privacy</h4>
The crawler log stores IP addresses and user-agent strings from Markdown requests. This data stays in your WordPress database and is never sent to external services. You can disable logging entirely, configure a retention period, or clear the log at any time from the settings page.

Serve Markdown

Serve Markdown versions of your posts and pages to AI agents and crawlers. Content negotiation, .md URLs, auto-discovery, and crawler logging.

kkarpieszuk

Kolibia Agent-Ready makes WordPress content available as Markdown so tools and agents do not have to scrape HTML. Behaviour follows common content negotiation patterns (similar in spirit to <a href="https://developers.cloudflare.com/fundamentals/reference/markdown-for-agents/" rel="nofollow ugc">Markdown for Agents</a> on Cloudflare): clients that prefer Markdown can request it explicitly.
What is included today
<ul>
<li>Markdown responses for singular content — For each built-in public post type registered by WordPress core (<code>post</code>, <code>page</code>, <code>attachment</code> on a typical site), a single-item URL can return <code>Content-Type: text/markdown; charset=UTF-8</code> instead of the themed HTML page.</li>
<li>Markdown for the posts index — When the site front page shows latest posts (WordPress “Your homepage displays: Your latest posts”), the homepage and other blog-index URLs return Markdown listing the same posts as the main loop.</li>
<li>Static or WooCommerce shop front page — When a page is set as the front page (including a WooCommerce shop page where WordPress reports <code>is_singular</code> as false), that page’s content is still served as Markdown on <code>/</code> when clients negotiate <code>text/markdown</code>.</li>
<li>When Markdown is served
<ul>
<li>The request includes an <code>Accept</code> header that lists <code>text/markdown</code> (requests with <code>text/markdown;q=0</code> are ignored).</li>
<li>Or the URL includes the query argument <code>output_format=md</code> (handy for browsers and manual testing).</li>
</ul>
</li>
<li>HTML to Markdown — Rendered post body is passed through WordPress’s <code>the_content</code> filters, then converted to Markdown using the <a href="https://github.com/thephpleague/html-to-markdown" rel="nofollow ugc">league/html-to-markdown</a> library. A short YAML front matter block includes <code>title</code> and <code>permalink</code>.</li>
<li>Password-protected content — If a password is required, the response is a small Markdown message instead of the full body.</li>
<li>Discovery in HTML — On the normal HTML view, the plugin prints <code><link rel="alternate" type="text/markdown" href="…?output_format=md" /></code> so clients can find the Markdown URL.</li>
<li>Admin shortcut — On the posts and pages list screens, a row action View as AI Agent opens the same Markdown URL (preview URLs are used for drafts where WordPress would show Preview).</li>
</ul>
Developer filters
<ul>
<li><code>kolibia_ar_markdown_post_types</code> — Defaults to all public, built-in post types from core; override to add custom post types or remove types (e.g. <code>attachment</code>).</li>
<li><code>kolibia_ar_post_markdown</code> — Filter the final Markdown string.</li>
<li><code>kolibia_ar_home_markdown</code> — Filter the Markdown for the posts index / homepage loop.</li>
<li><code>kolibia_ar_home_markdown_post_types</code> — Post types included in the homepage Markdown list (default: <code>post</code>).</li>
<li><code>kolibia_ar_markdown_password_required</code> — Filter the Markdown shown when a password is required.</li>
</ul>
Roadmap
Further “agent readiness” features (broader API discovery, additional formats) may be added in future releases.

Kolibia Agent-Ready

Kolibia Agent-Ready adds API discovery and Markdown responses so AI agents can use your WordPress content and services.

Serve Markdown vs Kolibia Agent-Ready