A proposed plain-text file that gives AI models a curated, easy-to-read map of your site's content.
llms.txt is a proposed standard: a Markdown file placed at the root of a website that offers large language models a clean, curated guide to the site's most important content. The idea is to give AI systems a concise, structured entry point, links to your key pages and documentation in plain text, rather than making them parse cluttered HTML.
It is an emerging convention, not an official rule, and adoption by major AI providers is still limited and evolving as of 2026. It is best understood as a low-cost, forward-looking signal, not a guaranteed ranking or citation lever.
You publish a Markdown file at /llms.txt listing your most valuable pages with short descriptions, optionally pointing to fuller plain-text versions of key documents. The format is human-readable and machine-friendly, so a model can quickly understand what your site covers and where the authoritative content lives.
As AI systems increasingly read the web to answer questions, giving them a clean map of your best content is a cheap hedge. Even where adoption is uncertain, the file is trivial to produce and does no harm, and it forces you to identify what your most important pages actually are.
The biggest mistake is over-promising results. llms.txt is not a confirmed ranking factor and major engines do not universally honour it yet. Treat it as good hygiene, not a substitute for accessible, high-quality content and a crawlable site.
llms.txt is a proposed standard — a Markdown file at the root of a site (/llms.txt) — that offers large language models a clean, curated guide to a site's most important content, in a format easy for them to consume. The motivation is that LLMs and AI answer engines increasingly read websites, but HTML pages are full of navigation, ads and markup that waste the model's limited context and obscure the substance. llms.txt is an attempt to do for AI what robots.txt and sitemaps did for crawlers: give machines a purpose-built, high-signal entry point to your content.
A typical llms.txt is a concise Markdown document: the site or project name, a short description, and curated links to the most important pages or documentation, often with brief annotations — sometimes paired with expanded "llms-full.txt" versions containing the actual clean content. It is an emerging, community-driven proposal rather than an official standard endorsed by the major AI companies, and adoption and whether the engines actually use it are still evolving. It is most established in developer-documentation contexts, where giving AI coding assistants clean, structured docs has clear value.
Because it is low-cost and low-risk, adding an llms.txt is a reasonable, forward-looking move — especially for documentation-heavy or content-rich sites that want to be well-represented as AI reading grows. It will not hurt anything (models that ignore it simply do not read it) and may help you be understood and cited more accurately if and as adoption spreads. The pragmatic stance is to treat it as a cheap hedge on where AI discovery is heading, not a guaranteed ranking or citation lever — do it if it is easy, keep it curated and current, but do not expect it to move the needle on its own yet.
Part of our defined terms knowledge graph — browse every entry in this branch.
Optimising content to be cited and surfaced by AI answer engines, not just ranked in blue links.
An AI model trained on vast text to predict and generate language, powering tools like ChatGPT.
Rendering a page's HTML on the server so it arrives ready to display, before JavaScript runs.
A longer, more specific search query with lower volume but clearer intent and less competition.
The ratio of a customer’s lifetime value to the cost of acquiring them.
Common questions
Straight answers on how this fits your marketing and build.
Still have questions? Talk to a specialist