llms.txt Explained: Purpose, Format, Example | OTL
ODD THEORY LABS

WE ARE AVAILABLE FOR DECEMBER PROJECTS

Code and text file open on a laptop screen
Odd Theory Labs
2026-08-20
AI Search
4 min read

llms.txt is a markdown file at a website's root that gives large language models a clean, structured summary of the site's purpose, key pages, and facts, helping AI systems understand and describe the business accurately.

The problem it solves

A modern website is a poor reading experience for a machine. Navigation, cookie banners, animation, duplicated boilerplate, and client-rendered sections all sit between a language model and the twenty sentences that actually describe your business.

llms.txt is a proposed answer: one markdown file at your root that says, plainly, what this site is, what it covers, and which pages matter.

What it is not

  • It is not a permissions mechanism. Access control stays in robots.txt.
  • It is not a ranking factor in traditional search.
  • It is not a place for keyword stuffing. Bloated files defeat the purpose.
  • It is not a substitute for structured data, sitemaps, or good page content.

The format

Plain markdown. A typical structure:

Heading one: the organisation or site name.

A blockquote summary: one to three sentences describing what the entity is and does. This is the most important line in the file, because it is the description most likely to be reused.

Optional context paragraphs: a short expansion for nuance.

Sections of links: grouped by purpose, each link followed by a short description of what that page contains.

An optional section of key facts: location, contact, services, sectors served.

Keep it under a few hundred lines. If it starts resembling a sitemap, you have gone too far.

A working example structure

A services business might publish:

  • A summary line stating category, location, and who it serves.
  • A section listing service areas, each linking to the relevant page with a one-line description.
  • A section of key pages: about, work, contact, insights.
  • A facts section: legal name, address, contact email, sectors, notable clients.
  • An optional section pointing to a fuller llms-full.txt for detail.

The discipline of writing it is itself valuable. Most businesses discover their own positioning is vague only when forced to state it in three sentences a machine will repeat.

Where to put it

At the site root, served as plain text or markdown with a suitable content type, reachable at yourdomain.com/llms.txt. Many sites also publish llms-full.txt with a longer reference version.

Link it from robots.txt if you like, and keep it in version control so it is updated alongside the site rather than forgotten.

What to include for a local business

  • Trading name and any common variations.
  • What you do, in one sentence, without jargon.
  • Cities or regions served.
  • Services, distinctly named.
  • Sectors or client types.
  • Contact email and physical address.
  • Links to the pages that expand each item.

What to leave out

  • Marketing superlatives. They are unusable and dilute the factual signal.
  • Private information, internal URLs, or anything behind authentication.
  • Pricing you are not willing to have quoted back to you.
  • Claims you cannot corroborate elsewhere, since inconsistency reduces trust.

Maintenance

Treat it as a living document:

  • Update when services, address, or positioning change.
  • Review quarterly alongside your about page, so the two never disagree.
  • Keep the facts identical to your structured data and business listings. Contradictions between sources are exactly what makes a model hedge or omit you.

Is it worth doing?

The cost is an hour. The downside is zero. The upside is that when a system tries to describe your business, there is a clean, self-authored, factually consistent description available in the easiest possible format.

Given how much time businesses spend on marketing assets nobody reads, publishing one file that machines might read is an unusually cheap bet.

Key takeaways

  • It is a summary for machines, not a permissions file — robots.txt still controls access.
  • Keep it factual, link the pages that matter, and maintain it like any other public asset.
  • Adoption is still emerging; the cost is trivial and the downside is nil.

Frequently asked questions

What is llms.txt?

A markdown file placed at the root of a website that summarises what the site is, what it offers, and where the important content lives, in a form language models can read cheaply and reliably.

Is llms.txt the same as robots.txt?

No. robots.txt controls which crawlers may access what. llms.txt provides a clean summary of the site's content for language models. They serve different purposes and should both exist.

Do AI companies actually use llms.txt?

Support is emerging rather than universal. It costs almost nothing to publish and maintain, and it doubles as a useful internal statement of what your site claims to be.

Related reading

llms.txtai searchtechnical seo

LET’s start
your project

IMPACTFUL CREATIVE GLOBAL SUPPORT

OFFICE

  • Ahmedabad

    Ahmedabad, Gujarat
    India