llms.txt is a proposed Markdown file that gives AI agents a short guide to a website. Google says its Search does not use it, and says it neither helps nor harms visibility. We could not verify public statements from OpenAI or Anthropic that their systems read it.
llms.txt is a proposal to put a file at /llms.txt that gives information to help agents use a website. The proposal page says it was published on September 3, 2024, was written by Jeremy Howard, and describes itself as v2 of the proposal. The page shows a modified date of August 10, 2026.
The idea is simple. A site gives brief background, guidance and links to detailed Markdown files. An agent can then find what it needs without processing full HTML pages. The file can sit at the site root or at a subpath, where it covers the pages under that path.
The proposal defines a Markdown file with a fixed order. Only the first heading is required.
Google's AI optimization guide for Search, last updated 2026-07-10, says you do not need to create new machine readable files, AI text files, markup or Markdown to appear in Google Search, including its generative AI capabilities. It says Google Search itself does not use llms.txt files. It adds that it is completely fine to create and maintain them, and that doing so will neither harm nor help your site's visibility or rankings in Google Search.
Earlier statements from Google's John Mueller point the same way. On Bluesky on June 17, 2025, he wrote: "FWIW no AI system currently uses llms.txt." That date comes from Search Engine Roundtable. Search Engine Journal reported on April 17, 2025 that Mueller wrote on Reddit that the file is comparable to the keywords meta tag, because it is what a site owner claims their site is about. These are individual statements, not product documentation.
The proposal page itself says thousands of sites publish an llms.txt file, that documentation platforms generate it automatically, and that OpenAI, Anthropic and Gemini publish llms.txt files for their developer docs. These are the proposal's own claims. We did not independently verify them.
Chrome's Lighthouse has an llms.txt audit. Its documentation, last updated 2026-05-05, calls llms.txt an emerging convention for a machine-readable summary of a site, designed for LLMs and AI agents. The audit only flags a server error when requesting the file. A missing file (404) is Not Applicable, because the file is optional for now.
We could not find a fetchable statement from OpenAI or Anthropic saying their crawlers or products read llms.txt to answer questions. Their crawler documentation, covered in our robots.txt reference, does not mention it. Whether they use it is unconfirmed.
Treat llms.txt as optional and low cost. It is not a ranking factor in Google Search, per Google. If you publish one, keep it short, accurate and in sync with your real site, and do not expect it to change AI citations. Spend effort first on crawlable, indexable pages with clear text content and correct robots.txt rules. If a tool or customer you care about says it reads the file, that is a reason to create it. Without that, there is no verified payoff.
Also keep the cost in view. The file is plain Markdown, so a short version takes little time to write. The risk is drift, where the file lists pages that have moved or been removed. If you publish one, add it to the checklist you already use when pages change, and test that it loads without a server error, since that is the one condition Lighthouse flags.
Sources were last checked on October 13, 2026. Documentation changes, so confirm any detail that matters to a decision against the source itself.
Indexive Research. (2026, October 9). What llms.txt is and who actually uses the file today. Indexive AI. https://indexive.ai/research/llms-txt-explained
Straight answers about how we work. Still have a question? The free audit is the fastest way to get one answered for your site.
Jeremy Howard, per the proposal page at llmstxt.org, which says it was published on September 3, 2024.
A free execution audit: your prioritized backlog, effort estimates, and what we'd ship first. Read-only access, no contract.