What is llms.txt?
llms.txt is a Markdown file at the root of a site (yourdomain.com/llms.txt) that tells large language models what the site is about and which pages matter, as a short curated list. It was proposed by Jeremy Howard (Answer.AI) at llmstxt.org in September 2024. It is a proposal, not an official web standard: unlike robots.txt it sets no access rules, and unlike sitemap.xml it does not list every URL — it is a short, human-curated site guide.
llms.txt format: required and optional parts
The proposal defines this order: (1) an H1 with the site or project name — the only required element; (2) a blockquote (>) with a short summary containing the key information needed to understand the rest; (3) optional detail paragraphs without headings; (4) H2 sections, each holding a list of - [Page name](https://full-url): short note entries; (5) a special ## Optional section for secondary resources that can be skipped when a shorter context is needed. Use absolute https:// URLs — a relative path such as /blog means nothing to a tool reading the file outside your site.
Example llms.txt file
A short example for an online store: "# Example Store" as the title, "> Online store selling handmade ceramics; product, shipping and returns information." as the summary, shipping, returns and about pages under "## Documentation", a category page and a care guide under "## Examples", and the blog plus sitemap under "## Optional". Six or seven links are enough. The "Fill with example" button builds exactly this file — swap in your own domain and pages. For a live, auto-generated file see yusufbayrak.net/llms.txt.
Which pages belong in llms.txt?
llms.txt is not a copy of your sitemap; its value comes from selection. Add the pages an AI assistant or agent should read to answer questions about you: product or service and pricing pages, shipping/returns/FAQ policies, documentation or guides, and your most-used tools. Leave out tag, filter, cart, login and pagination URLs. A practical ceiling for most sites is 20-50 links, each with a one-sentence note after the colon explaining what the page answers.
Does llms.txt actually work? The evidence
Honest answer: so far, only a little. Google has said it does not support llms.txt (Gary Illyes, July 2025); John Mueller compared it to the old keywords meta tag, and Google's 2026 guidance on AI features says such files are not needed to appear in AI Overviews or AI Mode. In Ahrefs' analysis of 137,000 domains, roughly 97% of llms.txt files received no requests in the month studied, and only a small share of requests came from named AI tools (led by GPTBot and Claude Code). SE Ranking's study of about 300,000 domains found no clear link between having the file and being cited by AI. On the other hand, Chrome's Lighthouse team added an experimental audit that checks for llms.txt as part of agentic-browsing readiness, and coding assistants and documentation tools can use it.
So why add one? Cost versus expectation
Writing llms.txt takes 10-15 minutes and breaks nothing: it cannot hurt rankings and does not change your robots.txt rules. Set expectations accordingly — it is not a traffic or citation lever, but a cheap aid that helps agents and developer tools understand your site quickly. Visibility in AI search still comes from the fundamentals: crawlable, indexed pages; content that answers the question in the first paragraph; correct structured data (JSON-LD); not blocking AI crawlers in robots.txt without reason; and links from other sites.
How to upload and test llms.txt
1) Save the file as UTF-8 plain text named exactly llms.txt. 2) Upload it to the site root; https://yourdomain.com/llms.txt should return status 200 as plain text (a response wrapped in your HTML template or redirected to a 404 page is useless). On WordPress place it in the root via FTP/file manager or use an SEO plugin's llms.txt feature; on platforms without root access check the method they support. 3) Make sure robots.txt does not block /llms.txt. 4) To measure, filter your server or CDN logs for /llms.txt requests by user agent (GPTBot, ClaudeBot, PerplexityBot, etc.); if no requests arrive, nobody is reading the file yet.
llms-full.txt and .md pages
llms.txt is only a list of links. Some documentation sites also publish llms-full.txt, which bundles the full content into one Markdown file, and the original proposal suggests serving a clean Markdown version of each page at the same URL with .md appended. These matter mostly for API and developer docs; for a typical store or services site, llms.txt is enough. Update the file whenever you add an important page — ideally generate it at build time, as this site does.