Selvakumar Duraipandian

Markdown for AI Agents is a lightweight WordPress plugin that enables HTTP content negotiation for your site’s content. When a client (like an AI agent or a custom script) requests a page with the <code>Accept: text/markdown</code> header, the plugin intercepts the request and returns a clean, structured Markdown representation of the post or page content.
This is ideal for AI crawlers, RAG (Retrieval-Augmented Generation) systems, and non-browser clients that prefer machine-friendly text over complex HTML.
Important note: This plugin is primarily a developer/integration tool. Human visitors browsing your site will never see any difference — the Markdown output is only served when explicitly requested via the <code>Accept: text/markdown</code> HTTP header. Normal browser requests always receive the standard HTML page.
Key Features:
<ul>
<li>Automatically detects <code>Accept: text/markdown</code> headers.</li>
<li>Converts HTML content to clean Markdown using the League HTMLToMarkdown library.</li>
<li>Strips away theme layout, navigation, headers, footers, and sidebars — serving only the main content.</li>
<li>Adds useful HTTP response headers: <code>Content-Type: text/markdown</code>, <code>Vary: Accept</code>, and <code>X-Markdown-Word-Count</code>.</li>
<li>Respects WordPress visibility rules and filters.</li>
<li>No configuration required — works out of the box for posts, pages, and custom post types.</li>
</ul>
<h3>How It Works</h3>
This plugin uses a standard web technique called HTTP content negotiation. The same URL on your site can serve different representations of the same content depending on what the client asks for:
<ul>
<li>A regular browser sends <code>Accept: text/html</code> → receives your normal HTML page.</li>
<li>An AI agent sends <code>Accept: text/markdown</code> → receives a clean Markdown version of the same page.</li>
</ul>
No extra URLs, no duplicate content, no configuration needed. The plugin hooks into WordPress’s <code>template_redirect</code> action, detects the <code>Accept</code> header, captures the rendered HTML, converts it to Markdown, and returns it with appropriate headers.
<h3>Why Markdown for AI Agents?</h3>
When building RAG (Retrieval-Augmented Generation) applications or AI pipelines that ingest web content, HTML is extremely noisy. A typical WordPress page contains thousands of tokens worth of HTML tags, inline styles, navigation menus, scripts, and layout markup — none of which carries meaning for an AI model.
Serving clean Markdown instead can reduce token consumption by up to 60%, which means:
<ul>
<li>Lower API costs — fewer tokens ingested when loading pages into vector stores or LLM pipelines.</li>
<li>Faster processing — less text for the model to parse, filter, and discard.</li>
<li>Better retrieval accuracy — higher signal-to-noise ratio improves the quality of RAG results.</li>
<li>Simpler pipelines — no need for custom HTML stripping logic on the client side; the plugin handles it server-side.</li>
</ul>
Any AI agent, crawler, or ingestion script that sends <code>Accept: text/markdown</code> in its request header will automatically receive the clean Markdown version — no extra URLs, no separate endpoints, no changes to your content workflow.

Markdown for AI Agents

Serve clean Markdown versions of WordPress content to AI agents using HTTP content negotiation.

Umay Ajans

Modern AI agents (ChatGPT, Claude, Perplexity, Gemini, etc.) work much better with Markdown than HTML. Umay AI Markdown inspects the incoming <code>Accept</code> header and, only when <code>text/markdown</code> is requested, intercepts the response and serves a clean, agent-friendly Markdown representation of the page.
Browsers, search engines, and any client that does not explicitly ask for Markdown receive the unchanged HTML response. There is no settings page, no cron job, and no external service call.
Key features
<ul>
<li>Zero configuration — install, activate, done.</li>
<li>Only triggers when <code>Accept: text/markdown</code> is present. Regular visitors and search engines are never affected.</li>
<li>Hybrid content extraction: uses <code>the_content</code> for posts/pages, falls back to a DOM-based extractor for archives, taxonomies, and the homepage.</li>
<li>Powered by the industry-standard <a href="https://github.com/thephpleague/html-to-markdown" rel="nofollow ugc">league/html-to-markdown</a> library.</li>
<li>Transient-cached for 12 hours per URL (sha256-keyed). Auto-invalidated on <code>save_post</code>, term edits, theme switches, and menu updates.</li>
<li>Built-in IP rate limiter (30 requests / minute by default) to mitigate abuse.</li>
<li>Strict input sanitization, header injection protection, libxml entity hardening (XXE-safe), and full WordPress Coding Standards compliance.</li>
<li>PSR-4 autoloaded, namespaced OOP code. No globals.</li>
<li>Sends <code>Vary: Accept</code>, <code>X-Robots-Tag: noindex</code>, and <code>X-Content-Type-Options: nosniff</code> on every Markdown response.</li>
</ul>
What gets sent to AI agents
Each Markdown response includes a YAML front-matter block with the page title, site name, canonical URL, and ISO-8601 generation timestamp, followed by the page body converted to Markdown. Navigation, footer, sidebars, scripts, styles, comment forms, related posts, and other page chrome are stripped before conversion.
Filters
Two filters are available for advanced customization:
<ul>
<li><code>umay_mdn_bypass</code> — Return <code>true</code> to skip Markdown handling for the current request.</li>
<li><code>umay_mdn_cache_ttl</code> — Override the default 12-hour cache lifetime (in seconds, minimum 60).</li>
<li><code>umay_mdn_rate_limit</code> — Override the default 30-requests-per-minute rate limit.</li>
<li><code>umay_mdn_converter_options</code> — Modify the league/html-to-markdown converter options array.</li>
</ul>

Umay AI Markdown

Serves clean Markdown to AI agents (ChatGPT, Claude, Perplexity) on Accept: text/markdown. HTML stays default for browsers. Zero-config.

Markdown for AI Agents vs Umay AI Markdown