llms.txt: a plain-English definition
llms.txt is a plain-text file, written in Markdown and placed at the root of a website (yoursite.com/llms.txt), that gives AI language models a clean, curated map of the site’s most important content. Proposed by Jeremy Howard of Answer.AI in September 2024, it aims to help models find and understand key pages without wading through HTML, scripts and navigation.
What llms.txt does
Modern web pages are built for humans — menus, scripts, ads, styling — but AI models read the raw code and operate within limited context windows. An llms.txt file strips that back to a simple Markdown briefing: an H1 with your brand name (the entity anchor), a short blockquote summary describing what you are and who you serve, and curated links to your most important pages. It is a navigation aid, not a blocking tool — it cannot restrict crawlers the way robots.txt can.
The most important line is the blockquote summary: written factually and specifically, it’s what a model uses to understand your identity. Marketing fluff there tells the model nothing.
Do you actually need one?
Here’s the honest position. llms.txt is a community convention, not backed by W3C, IETF or any standards body, and as of early 2026 no major AI provider — OpenAI, Google, Anthropic, Meta — has publicly committed to reading it in production. GPTBot fetches it occasionally; adoption is inconsistent.
That said, it costs almost nothing to add, makes your site more legible to anything that does read it, and carries no downside if ignored. The sensible view: ship one as a low-cost hedge, keep it accurate, but don’t treat it as a guaranteed visibility lever. It complements, rather than replaces, structured data and genuine content quality.
How to write one
Use valid Markdown, served as plain text, literally named llms.txt (lowercase), at the domain root. Start with a single H1 (your brand), follow immediately with a one-to-three-sentence blockquote summary, then group your most important pages under H2 sections as annotated links. Keep it curated — it’s a briefing, not a full sitemap — and review it quarterly so it reflects your current positioning.
llms.txt, answered.
What is llms.txt?
llms.txt is a Markdown text file placed at a website's root that gives AI models a clean, curated map of the site's most important content. Proposed by Jeremy Howard of Answer.AI in September 2024, it helps models find and understand key pages without processing HTML clutter. It's a navigation aid, not a crawler-blocking tool.
Do AI models actually read llms.txt?
Inconsistently. As of early 2026, llms.txt is a community convention with no standards-body backing, and no major AI provider — OpenAI, Google, Anthropic, Meta — has publicly committed to reading it in production. GPTBot fetches it occasionally. It's best viewed as a low-cost hedge rather than a guaranteed visibility lever.
Is llms.txt the same as robots.txt?
No. Both are plain-text files at a site's root, but robots.txt tells crawlers what they may or may not access, while llms.txt tells AI models which content matters most and how it's organised. llms.txt cannot block or restrict any crawler — it's a navigation and summary aid, not an access-control tool.
Do you need llms.txt to appear in AI answers?
No. Google has explicitly said llms.txt is not required for AI Overviews, and adoption across providers is inconsistent. It's a low-cost, low-risk addition that may help models that read it, but it isn't a prerequisite for AI visibility. Content quality, structure and entity authority matter more.
What goes in an llms.txt file?
A single H1 with your brand name, a short blockquote summary (the most important line — a factual description of what you are and who you serve), optional detail paragraphs, and H2 sections listing your most important pages as annotated links. It should be curated, not a dump of every URL.
Want to put llms.txt into practice?
Cllimber is a curated, industry-specific source built to be cited by AI — for software and service providers.