llms.txt
llms.txt is a small text file you put on your website that hands AI tools a curated map of your most important pages — like a sitemap, but for language models.
What it is
llms.txt is a simple proposed standard: a plain-text file at yoursite.com/llms.txt that lists the canonical pages an AI tool should read first when trying to understand your business. Think of it as a short, opinionated guide for language models — not a list of every page (that's sitemap.xml), but a curated index of the few pages that actually represent who you are and what you offer.
Adoption is still mixed in 2026 — not every AI tool reads llms.txt yet — but the cost of adding one is low (often a few minutes), and the upside is real for tools that do read it. Even AI tools that ignore the file directly still benefit indirectly: writing an llms.txt forces you to make your canonical content easy to find, which helps regular crawlers too.
Why it matters for your business
AI tools that hit your site for the first time face a problem: which pages matter? llms.txt solves that. Instead of letting the model guess from your nav, footer, or random internal links, you hand it a short list: "here's our main services page, here's our service-area page, here's our reviews page, here's our FAQ." That guides what gets summarized when an AI is asked about your business — and reduces the chance it pulls from a stale or outdated page.
Example
A multi-location HVAC company has a sprawling website with dozens of city and service pages, plus a blog, careers section, and old promotions. They added a 12-line llms.txt listing their canonical service pages, the master service-area page, the reviews page, and the FAQ. Within a few weeks, AI assistants started consistently citing the right pages when summarizing the company — instead of randomly surfacing a two-year-old promo page that was still live.
Related
- Guide: LLMs.txt Guide, AI Discovery Surfaces (AEO/GEO playbook)
- Related terms: AEO, GEO, Extractability, Schema / structured data