gregrandall

AI agents, LLMs, and crawlers have to wade through navigation bars, sidebars, ads, and comment forms to reach the content they want, and every element costs tokens. <a href="https://blog.cloudflare.com/markdown-for-agents/" rel="nofollow ugc">Cloudflare measured</a> an 80% reduction in token usage when converting a blog post from HTML to Markdown (16,180 tokens down to 3,150).
Botkibble adds a Markdown endpoint to every published post and page.
Cloudflare offers <a href="https://developers.cloudflare.com/fundamentals/reference/markdown-for-agents/" rel="nofollow ugc">Markdown for Agents</a> at the CDN edge on Pro, Business, and Enterprise plans. Botkibble does the same thing (for free) at the origin, so it works on any host.
<a href="https://github.com/greg-randall/botkibble" rel="nofollow ugc">GitHub Repository</a>
Three ways to request Markdown:
<ul>
<li><code>.md</code> suffix: append <code>.md</code> to any post or page URL (e.g. <code>example.com/my-post.md</code>)</li>
<li>Query parameter: add <code>?format=markdown</code> to any post or page URL</li>
<li>Content negotiation: send <code>Accept: text/markdown</code> in the request header</li>
</ul>
What’s in every response:
<ul>
<li>Structured metadata header with title, date, categories, tags, word count, character count, and estimated token count (in YAML frontmatter format, readable by any AI agent)</li>
<li>Clean Markdown converted from fully-rendered post HTML (shortcodes run, filters applied)</li>
<li><code>Content-Type: text/markdown</code> and <code>Vary: Accept</code> response headers</li>
<li><code>Content-Signal</code> header for AI signal declaration — defaults to <code>ai-train=no, search=yes, ai-input=yes</code> — see <a href="https://contentsignals.org/" rel="nofollow ugc">contentsignals.org</a></li>
<li><code>X-Markdown-Tokens</code> header with estimated token count</li>
<li>Discovery via <code><link rel="alternate"></code> in the HTML head and <code>Link</code> HTTP header</li>
<li>Automatic cache invalidation when a post is updated or deleted</li>
</ul>
Performance:
Botkibble writes Markdown to disk on the first request, then serves it as a static file. A built-in Fast-Path serves cached files during WordPress’s <code>init</code> hook, before the main database query runs. No extra configuration needed.
Add a web server rewrite rule and Botkibble bypasses PHP entirely, serving <code>.md</code> files the same way a server would serve an image or CSS file:
 Method 
 Avg. response time
 Standard HTML 
 0.97s
 Markdown (cold, first request) 
 0.95s
 Markdown (cached, PHP Fast-Path) 
 0.87s
 Markdown (Nginx/Apache direct) 
 0.11s
Serving directly from disk is 88% faster than a full WordPress page load. See the Performance section below for Nginx and Apache configuration.
Security:
<ul>
<li>Drafts, private posts, and password-protected content return <code>403 Forbidden</code></li>
<li>Rate limits cache-miss regenerations (20/min by default) to mitigate DoS abuse</li>
<li><code>X-Robots-Tag: noindex</code> keeps Markdown versions out of search results</li>
<li><code>Link: rel="canonical"</code> points search engines back to the HTML version</li>
</ul>
Cache variants (optional):
You can persist alternate cached representations by adding <code>?botkibble_variant=slim</code> (or any other variant name). 
Variant caches are stored under:
<pre><code>/wp-content/uploads/botkibble/_v/<variant>/<slug>.md
</code></pre>
What it does NOT do:
<ul>
<li>Expose drafts, private posts, or password-protected content</li>
<li>Serve non-post/page content types by default</li>
<li>Require any configuration. Activate and it works.</li>
</ul>
<h3>Why Markdown?</h3>
HTML is expensive for AI systems to process. <a href="https://blog.cloudflare.com/markdown-for-agents/" rel="nofollow ugc">Cloudflare measured</a> an 80% reduction in token usage when converting a blog post from HTML to Markdown (16,180 tokens down to 3,150).
Cloudflare now offers <a href="https://developers.cloudflare.com/fundamentals/reference/markdown-for-agents/" rel="nofollow ugc">Markdown for Agents</a> at the CDN edge via the <code>Accept: text/markdown</code> header, available on Pro, Business, and Enterprise plans.
This plugin does the same thing at the origin, so it works on any host. It also adds <code>.md</code> suffix URLs, <code>?format=markdown</code> query parameters, YAML frontmatter, static file caching, and server-level offloading.
If you use Cloudflare, both share the same <code>Accept: text/markdown</code> header, <code>Content-Signal</code> headers, and <code>X-Markdown-Tokens</code> response headers.
Cloudflare currently defaults to <code>Content-Signal: ai-train=yes, search=yes, ai-input=yes</code> with no way to change it. Botkibble defaults to <code>ai-train=no</code> and lets you override the full signal per site via the <code>botkibble_content_signal</code> filter.
<h3>Performance & Static Offloading</h3>
This plugin supports static file offloading by writing Markdown content to <code>/wp-content/uploads/botkibble/</code>.
<h3>Nginx Configuration</h3>
To bypass PHP entirely and have Nginx serve the files (including variants) directly:
<pre><code># Variants
location ~* ^/(_v/[^/]+/.+)\.md$ {
 default_type text/markdown;
 try_files /wp-content/uploads/botkibble/$1.md /index.php?$args;
}

# Default
location ~* ^/(.+)\.md$ {
 default_type text/markdown;
 try_files /wp-content/uploads/botkibble/$1.md /index.php?$args;
}
</code></pre>
<h3>Apache (.htaccess)</h3>
Add this to your <code>.htaccess</code> before the WordPress rules:
<pre><code>RewriteEngine On
# Variants
RewriteCond %{DOCUMENT_ROOT}/wp-content/uploads/botkibble/_v/$1/$2.md -f
RewriteRule ^_v/([^/]+)/(.+)\.md$ /wp-content/uploads/botkibble/_v/$1/$2.md [L,T=text/markdown]

# Default
RewriteCond %{DOCUMENT_ROOT}/wp-content/uploads/botkibble/$1.md -f
RewriteRule ^(.*)\.md$ /wp-content/uploads/botkibble/$1.md [L,T=text/markdown]
</code></pre>
Even without these rules, the plugin uses a “Fast-Path” that serves cached files from PHP before the main database query is executed.
<h3>Credits</h3>
We thank Cristi Constantin (https://github.com/cristi-constantin) for contributing cache variants, URL and SEO improvements, and fixing important bugs.

Botkibble

Serves every published post and page as Markdown for AI agents and crawlers. No configuration, no API keys. Activate and it works.

Jon Henshaw

Coywolf Robots.txt Manager turns robots.txt into a managed list of rules you build from the WordPress admin — no FTP, no hand-editing. It ships the full catalog of bots from the Cloudflare Bot Directory, grouped by category (AI Crawler, Search Engine Crawler, Advertising, Security, and more), so you can disallow whole categories of unwanted bots with a click.
<ul>
<li>Rules table — every directive as a row with a name and plain-English description, with bulk delete and pagination. Identical directives are consolidated.</li>
<li>Direct editor — a Robots.txt page shows the current file in an editable box; saving re-parses and optimizes your hand edits into managed rules. A copy button puts the whole robots.txt on your clipboard.</li>
<li>Guided rule editor — choose a Rule Type (Disallow or Allow) and a Rule Path (folder, prefix, file type, query string, allow-exception, and more); the editor builds the correct directive, with auto-filled name and description, plus a URL testing tool.</li>
<li>Conflict checking — every add or edit is checked against existing rules; if it duplicates or contradicts one, the editor explains the clash and won’t save until it’s resolved.</li>
<li>Bot catalog — every bot from the Cloudflare Bot Directory, plus a curated set of well-known AI crawlers it doesn’t list yet (ClaudeBot, CCBot, Google-Extended, Applebot-Extended, and more), grouped by category, with select-all per category. The catalog ships with the plugin and is refreshed through plugin updates.</li>
<li>Curated robots.txt tokens — the exact User-agent token each bot obeys is curated for all 600+ bots (the directory’s own match data is often a URL, operator name, or generic word), so rules block the right token and imported robots.txt files match the right bot.</li>
<li>Commented output — each rule is written as its own section prefixed with a comment carrying its name and description; a rule’s bots are stacked under User-agent: lines with the directive written once.</li>
<li>Physical or virtual — write a real robots.txt in your site root, or let WordPress serve it virtually. On install, an existing physical file is detected, managed, and imported automatically.</li>
<li>Backup & restore — your original robots.txt is captured before the plugin reads, parses, or rewrites anything; on deactivation it asks whether to restore that original or keep the rules it built (markers removed).</li>
<li>Import & export — on the Import / Export Rules page, download all rules and settings preferences as one JSON file, then import it to restore them or copy the setup to another site. Import replaces the current rules; sitemap URLs aren’t included (they’re site-specific).</li>
<li>Optional comments — the per-rule description comments can be excluded from robots.txt under Settings → Rule comments, and added back any time.</li>
<li>Role-based access — under Settings → Access, choose which user roles can see and use the plugin; Administrator is always allowed and can’t be locked out. Only administrators can change the access list and the serving mode.</li>
<li>Cleans up on import — fixes misspelled directives, missing colons, and missing leading slashes; consolidates duplicate user-agent groups and resolves Allow/Disallow conflicts; strips deprecated/unsupported directives (Crawl-delay, Noindex, Nofollow, Host, Request-rate, Visit-time); removes render-blocking /.css and /.js blocks and the overly broad /wp-content/; fixes malformed wildcards; and makes Sitemap links absolute and de-duplicated, dropping any that point at a different domain (also handles a UTF-8 BOM and Google’s 500 KiB limit).</li>
<li>XML sitemaps — add sitemap URLs under Settings → XML sitemaps (written at the end of robots.txt), or exclude sitemap links entirely.</li>
<li>Owns the virtual robots.txt — in virtual mode the plugin serves the whole file. WordPress’s own default rules (Disallow: /wp-admin/) are imported and reconciled into your managed rules when you activate or switch to virtual (WordPress’s rules win, so its “*” rule absorbs a matching per-bot rule), and WordPress stops adding a duplicate block on top. Other SEO plugins’ additions (e.g. Yoast) are excluded too. Physical mode is served by WordPress directly.</li>
</ul>
<h3>Conflict resolution</h3>
When rules are imported, or you hand-edit the Robots.txt page and save, the plugin re-parses everything and resolves conflicts automatically with opinionated, deterministic decisions:
<ul>
<li>Identical directives that differ only by user-agent are consolidated into one rule listing every agent; duplicate User-agent groups are merged case-insensitively.</li>
<li>When an agent both allows and disallows the same path, the later rule wins, and a blanket Disallow: / is dropped for an agent that has more specific rules.</li>
<li>If a directive applies to * as well as named bots for the same path, it collapses to just * (the wildcard already covers them).</li>
<li>In virtual mode WordPress’s own User-agent: * base rules (the /wp-admin/ block) win and absorb any matching per-bot rule, and are served once instead of duplicated on top.</li>
<li>Deprecated/unsupported directives are removed (Crawl-delay, Noindex, Nofollow, Host, Request-rate, Visit-time).</li>
<li>SEO-harmful blocks are removed (render-blocking /.css and /.js, and the overly broad /wp-content/).</li>
<li>Syntax is repaired (misspelled directives, a missing colon, a missing leading slash, malformed wildcards).</li>
<li>Sitemap links are made absolute and de-duplicated (preferring https); cross-domain ones are dropped; a UTF-8 BOM is stripped and Google’s 500 KiB limit is honored.</li>
</ul>
<h3>Bot tokens</h3>
Blocking a bot only works if robots.txt uses the exact User-agent token that bot obeys. The bundled Cloudflare Bot Directory isn’t reliable for this — its match data is often a URL, the operator’s name, or a generic word like “Bot” — so every one of the 600+ bots has a curated robots.txt token, researched against each operator’s documentation and the product token in its real user-agent (e.g. Toutiao → Bytespider, Internet Archive → archive.org_bot, BrightEdge Bot → BrightEdge). Each entry records the variant spellings that should map to it, so importing an existing robots.txt matches the right catalog bot. The tokens are bundled with the plugin and kept current through plugin updates. Bots that publish no robots.txt-targetable token (e.g. “signed agents” that identify cryptographically) are left out of the list, since robots.txt can’t target them.

Coywolf Robots.txt Manager

Manage robots.txt from the WordPress admin: build rules, block bots by category, and serve a physical or virtual robots.txt.

Botkibble vs Coywolf Robots.txt Manager