What Is llms.txt? A Practical Guide
llms.txt is a Markdown file at your site root that tells AI models what your site is and which pages matter. Here is the format, a copy-ready example, an honest read on whether it works, and how it fits into AI visibility overall.
The Problem It Solves
When a language model needs to understand your business, it faces a website built for humans: navigation menus, cookie banners, marketing copy, and hundreds of pages of varying importance. It has a limited context window and no reliable way to know that your services page matters more than your 2019 holiday-hours post.
llms.txt is the proposed fix. Published at yoursite.com/llms.txt, it is a short Markdown document that states what your site is, who it serves, and which URLs are worth reading - in clean text with no markup to wade through. Jeremy Howard of Answer.AI proposed the format in September 2024.
llms.txt vs robots.txt vs sitemap.xml
| File | Audience | Purpose | Status |
|---|---|---|---|
| robots.txt | Crawlers | What may and may not be fetched | Widely honored |
| sitemap.xml | Search engines | Every URL that exists, for discovery | Established standard |
| llms.txt | Language models | Which URLs matter, and what the site is | Community proposal |
The distinction that matters: a sitemap is exhaustive, llms.txt is curated. A sitemap lists everything; llms.txt is an argument about what deserves attention.
The Format
H1 - your site name
A single top-level heading with the name of the site or project. This is the only required element of the format.
Blockquote summary
One italicized or blockquoted sentence stating what the site is and who it serves. This is the line a model is most likely to reuse verbatim, so write it deliberately.
Context paragraphs
Optional plain prose adding detail a model would otherwise have to infer - pricing model, geography served, what you explicitly do not do.
H2 link sections
Grouped bulleted lists of your important URLs, each with a short description of what the page contains. Group them logically: core pages, guides, product.
Optional section
A final H2 literally titled 'Optional' holds secondary links a model may skip when it is short on context. Everything above it is treated as primary.
A Complete Example
Copy this, replace the details with your own, and save it as llms.txt in your site's public root folder.
# Acme Plumbing
> Licensed emergency and residential plumbing in Spokane, WA. Same-day service, flat-rate pricing.
Family owned since 1998. We serve Spokane County only and do not take on
new commercial construction contracts.
## Core pages
- [Services](https://example.com/services): Full list of residential and emergency plumbing services with flat-rate pricing.
- [Service area](https://example.com/service-area): The Spokane County cities and neighborhoods we cover.
- [Pricing](https://example.com/pricing): Flat-rate pricing table and after-hours rates.
## Guides
- [What to do in a burst pipe emergency](https://example.com/guides/burst-pipe): Step-by-step shutoff instructions before a plumber arrives.
## Optional
- [About](https://example.com/about): Company history and licensing details.
- [Contact](https://example.com/contact): Phone, email, and hours.Rules Worth Following
- Use absolute URLs, not relative paths - the file may be read outside your domain's context.
- Keep it short. Twenty well-chosen links beat two hundred; curation is the whole point.
- Write the summary line for a reader who has never heard of you and will read nothing else.
- State what you do not do. Negative constraints prevent a model from recommending you for the wrong thing.
- Serve it as text/plain or text/markdown, not HTML.
- Update it when your key pages change - a stale file is worse than none.
Does It Actually Work?
Honestly: unproven. No major AI provider has publicly confirmed that it reads llms.txt, and there is no independent study demonstrating a traffic or citation lift from adding one. Anyone selling you guaranteed AI visibility on the strength of this file is ahead of the evidence.
It is still worth doing, for three reasons. It costs twenty minutes. It cannot damage your existing search performance, because it is an additional file rather than a change to anything Google already reads. And if adoption grows, the sites that already have a clean, accurate one are positioned for it.
What it will not do is compensate for weak fundamentals. If your business name and phone number differ across three directories, if your site has no structured data, and if your competitors list attributes you do not, a curated link file will not rescue you. Those are the signals that actually decide whether an AI answer names you.
Where It Fits
llms.txt is one line item in a broader AI visibility program. The larger work is entity consistency, structured data coverage, review sentiment, and competitive attribute parity - the things a model weighs when deciding whom to recommend.
If you want to know where you stand on those, an AI visibility audit scores each of them and ranks the fixes by impact. You can read a full sample report free with no signup. For the bigger picture, see AEO vs SEO.
llms.txt: Common Questions
What is llms.txt?
llms.txt is a plain Markdown file published at the root of your website - yoursite.com/llms.txt - that tells large language models what your site is about and which pages matter most. It was proposed by Jeremy Howard of Answer.AI in September 2024. Think of it as a curated map written for a model with a limited context window, rather than a crawl directive like robots.txt.
Is llms.txt an official standard?
No. It is a community proposal, not a W3C or IETF standard, and no major AI provider has publicly committed to reading it. Adopting it is a low-cost bet: the file takes minutes to write, cannot hurt your existing SEO, and positions you if adoption grows. Treat anyone promising guaranteed AI traffic from llms.txt with skepticism.
How is llms.txt different from robots.txt?
robots.txt is a permission file - it tells crawlers which paths they may not fetch, and it is widely honored. llms.txt is a comprehension file - it tells a model which of your pages are worth reading and what your site is for. robots.txt restricts; llms.txt recommends. They serve different purposes and you should have both.
How do I create an llms.txt file?
Write a Markdown file with a single H1 for your site name, an italicized one-line summary, optional context paragraphs, then H2 sections containing bulleted lists of links in the format: dash, bracketed page title, parenthesized absolute URL, colon, short description. Save it as llms.txt in your site's public root so it serves at yoursite.com/llms.txt.
Does llms.txt help my SEO?
Not directly - Google has not said it uses the file as a ranking signal. Its value is in answer engine optimization: making your key pages and your positioning unambiguous to a model that may only read a fraction of your site. The exercise of writing one also forces useful clarity about which pages actually matter.
What is llms-full.txt?
An optional companion file that contains the full text content of your documentation or key pages in one Markdown document, rather than just links. It is mainly used by documentation-heavy sites so a model can ingest everything in a single fetch. Most local business and small product sites only need llms.txt.