lw-seo-machine-readable-content

LW SEO machine-readable content

Treat the sitemap, llms files, Markdown endpoint, and Content Signals as separate outputs with one shared content eligibility boundary. Test what an anonymous visitor receives because llms caches are built logged out and served to everyone.

Output map

Output	Default	Main contract
/sitemap.xml	enabled	index plus per-type XML files
virtual /robots.txt additions	enabled	sitemap, llms links, crawler rules, Content Signal
/llms.txt	enabled	one section per enabled public post type
/llms-full.txt	disabled	concatenated Markdown, capped at 1 MiB
/{object}/md and /markdown	enabled for eligible objects	explicit Markdown response
?format=md	enabled for eligible objects	explicit Markdown response
Accept: text/markdown	negotiated	singular HTML URL can return Markdown

Activation registers rewrite rules for enabled virtual endpoints. Toggling sitemap, llms.txt, or llms-full.txt schedules a soft rewrite flush for the next request.

Apply the shared eligibility gate

A post is eligible only when it is published, has no password, belongs to a viewable post type, is not globally or individually noindex, and passes lw_seo_post_is_eligible. AI output also requires ai-input not to resolve to no.

Use the restrict-only filter for membership, LMS, intranet, or paid content:

add_filter(
	'lw_seo_post_is_eligible',
	static function ( bool $eligible, WP_Post $post ): bool {
		if ( my_plugin_requires_entitlement( $post->ID ) ) {
			return false;
		}

		return $eligible;
	},
	10,
	2
);

This removes an otherwise eligible object from the sitemap, llms.txt, llms-full.txt, and Markdown endpoint. It cannot make an ineligible object public.

Do not rely only on the_content masking. llms-full.txt renders entries outside the normal singular main query. Mask descriptions with get_the_excerpt or lw_seo_meta_description, and exclude objects at the eligibility gate when no machine-readable trace should remain.