Create and manage llms.txt

The LLMs.txt module publishes a curated Markdown index at the exact root URL /llms.txt. It helps AI agents find the small set of resources that best explains the site; the XML sitemap remains the exhaustive URL inventory.

Enable it under Settings → POM Tweaks → SEO & LLMs → LLMs.txt.

Structure of the generated file

The output contains:

1. one H1 using Site name;
2. an optional introductory blockquote and plain-language context;
3. the human Last reviewed date when supplied;
4. H2 sections containing selected resources;
5. the exact ## Optional section for secondary resources and the sitemap.

WordPress entries use their public title and canonical permalink. A manual excerpt becomes the description; otherwise the module derives plain text from stored content without executing shortcodes. Term and archive descriptions are handled similarly.

Eligibility safeguards

Only explicitly selected public resources are candidates. During generation the module rejects content that is no longer published or publicly eligible, including password-protected and recognized noindex or alternate-canonical entries. Selected empty terms are also omitted.

Rank Math integration uses its indexability decision when available. Common noindex and canonical metadata from Rank Math, Yoast SEO, and SEOPress are recognized.

Static response behavior

The generated Markdown is cached so ordinary /llms.txt requests do not repeat all content queries. Relevant settings, selected content, taxonomy, SEO metadata, site identity, permalink, and POM Cache purge events invalidate that cached output.

The endpoint supports GET, HEAD, conditional requests, and the exact file-like URL. A physical llms.txt in the web root can take precedence before WordPress runs.

Configure and verify

1. Choose content for llms.txt.
2. Validate the public llms.txt response.

Disabling the module removes its endpoint hooks. Another plugin or a physical file can still answer the same URL.