What is llms.txt, and does your WordPress site need one?
llms.txt gets a lot of attention as a way to be seen by AI assistants. It is useful, cheap and easy to overrate.
Short answer
llms.txt is a proposed convention: a Markdown file at your site's root that gives AI tools a short summary and a list of your most useful pages. Some AI tools read it, and Google says Google Search ignores it. It is cheap to add and unlikely to hurt, but clear, crawlable pages and structured data matter more.
On this page5 sections
What llms.txt is
llms.txt is a proposed convention, published at llmstxt.org in September 2024, for a Markdown file at the root of a website: /llms.txt. It gives language models a short, readable map of the site: a title, a one-line summary, and lists of links to the most useful pages with a short note on each. The idea is that a model, or a tool working on someone’s behalf, can read that file instead of finding its way through menus, scripts and layout.
The proposal also suggests offering a clean Markdown version of a page at the same address with .md added. A common companion file, llms-full.txt, goes further and puts the site’s main content into a single Markdown document. A minimal llms.txt looks like this:
# Example Studio
> Independent web studio for small businesses.
## Services
- [Website fixes](https://example.com/services/fixes.md): Bugs, layout and speed
What it can and cannot do
It is worth being plain about the status. llms.txt is not a web standard, and no major search engine has confirmed that it uses the file for ranking or for its AI answers. Google’s guidance is explicit: for Google Search you can ignore AI text files such as llms.txt, and there are no additional requirements to appear in its AI features beyond the usual ones for being indexed. Some AI tools and coding assistants do read llms.txt when they find it, especially on documentation sites.
What it can do:
- give an assistant that fetches your site a clean summary and a short list of the pages that matter;
- offer text without navigation, cookie banners and markup, which is easier to quote accurately;
- make you decide which pages actually describe your business, which is a useful exercise in itself.
What it cannot do: make a model cite you, fix thin pages, or replace being crawlable and indexed. Treat it as a low-cost addition, not a strategy. The wider practice is often called generative engine optimization, and most of it is ordinary good publishing.
The basics that matter more
- Clear pages. One topic per page, a heading that says what the page is, and a direct answer near the top. People and models both quote the sentence that answers the question.
- Crawlable content. Important text should be in the HTML, not loaded only after a click or locked inside images. Check that your robots.txt does not block crawlers you want to read the site.
- Structured data. Structured data, usually written as JSON-LD, states facts such as your business name, address, services and opening hours in a form machines can read without guessing. Google documents which types it supports and how it uses them.
- Consistent facts. The same name, address, phone number and service list on your site and on your profiles elsewhere.
- Real expertise. Pages written by people who do the work, with specifics a generic page would not have.
How our free plugin handles it
We built LLM & GEO Optimizer for WordPress sites and published it free on GitHub. It generates three things:
/llms.txt: a site index grouped by post type and primary taxonomy, with titles, links and descriptions;/llms-full.txt: the content in Markdown, up to a size limit you set (100,000 characters by default);- a Markdown version of any page at
/any-page.mdor with?format=md, with YAML front matter for title, description, URL, date and taxonomy terms.
Details that matter in practice:
- It reads content the way WordPress renders it, so Gutenberg, the Classic Editor and shortcode-based builders work. If a page comes out thin, it falls back to Elementor’s rendering and then to Advanced Custom Fields or Secure Custom Fields.
- Descriptions come from Rank Math, Yoast SEO or All in One SEO, if you use one of them.
- Every Markdown endpoint and llms.txt file is sent with an
X-Robots-Tagnoindex header, so search engines are told not to index them alongside your real pages. - The cache is rebuilt when content is published, updated or deleted.
- A robots.txt check shows which AI crawlers, such as GPTBot, ClaudeBot and PerplexityBot, your current rules allow or block.
- On TranslatePress sites, each language gets its own llms.txt, llms-full.txt and Markdown documents.
It does not touch your HTML pages, meta tags, sitemaps or schema. After a deploy, the files can be rebuilt with WP-CLI:
wp llm-geo warm --flush
So, do you need one?
If your pages are clear, crawlable and well described, adding llms.txt is cheap and unlikely to hurt, so it is reasonable to have one. If they are not, fix that first: a summary file cannot point to answers that are not on the site. We cover both sides in our SEO and AI search work. If you would like a second opinion on where your site stands, get in touch.
Requirements: LLM & GEO Optimizer on WordPress 5.8+, PHP 7.4+
Key takeaways
- llms.txt is a proposal published at llmstxt.org, not a web standard that search engines are confirmed to use.
- It can give AI tools a clean summary and clean Markdown text, but it cannot make a model cite you.
- Clear pages, crawlable HTML, structured data and consistent facts matter more than any summary file.
- Our free LLM & GEO Optimizer plugin generates llms.txt, llms-full.txt and per-page Markdown, all marked noindex.
