llms.txt: What It Is and Whether Your Site Needs One

llms.txt in plain terms: what the file is, what it can and cannot do for AI visibility, and a copy-ready implementation that costs ten minutes — no magic, one real problem solved.

1,258 words · 6 min read · 9 min listen

plate 31 — rendered for this recording

read by jamie mckaye — his own voice, via his voice model. not a studio take.

00:00 / --:--

Every few months a new file format promises to transform your search rankings. Most are noise. llms.txt is different — not because it's a magic lever, but because it costs almost nothing to implement and solves a real problem: AI agents struggle to navigate websites built for humans, not machines.

Here is what it actually is, what it does, and whether your business should bother.

What Is llms.txt?

llms.txt is a proposed open standard, published at llmstxt.org, for a plain-text Markdown file placed at the root of your website (/llms.txt). Think of it as a structured guide — written for AI language models rather than web crawlers — that tells them what your site is, what it covers, and where to find the most important content.

The format is deliberately simple:

One H1 heading — the site or business name
A blockquote — a concise summary of what the site does
Optional H2 sections — curated lists of links to key pages, with brief descriptions

An optional companion file, llms-full.txt, goes further by inlining the actual content of those pages rather than just linking to them. This is useful when you want an AI to be able to read your site's substance in a single request, without crawling dozens of individual URLs.

The spec was introduced in late 2024 by Jeremy Howard (fast.ai) and has since been adopted by a growing number of developer tools and documentation sites.

What Problem Does It Solve?

When an AI agent — whether that's a research assistant, an autonomous agent, or an AI-powered search feature — lands on your website, it faces the same challenge a human does: too much navigational chrome, too many JavaScript-rendered components, too many links. It has to guess which content matters.

llms.txt gives the AI a shortcut: a curated, human-authored map of what is worth reading. For sites with large content libraries — software documentation, legal resources, technical blogs, product catalogues — this can meaningfully improve how thoroughly and accurately AI tools understand your content.

For a five-page brochure site selling window cleaning in Cheltenham, the benefit is more marginal.

What llms.txt Does NOT Do

Let's be direct about what the evidence actually shows as of mid-2026.

No major AI vendor has officially confirmed that llms.txt affects ranking or citation. OpenAI, Anthropic, Google, and Perplexity have not published documentation stating that they consume llms.txt files when deciding what to surface in AI Overviews, ChatGPT answers, or Claude's responses.

Google has explicitly stated that llms.txt is not required for Google Search and that it does not use the file for standard search ranking.

That said, Google's Lighthouse audit tool added an "Agentic Browsing" category in May 2026, which checks for the presence and validity of llms.txt. That is a signal: Google's developer tooling is treating it as a best-practice indicator, even if the ranking impact is unproven.

The honest framing: llms.txt is a low-cost, forward-looking signal. It does not replace the fundamentals — crawler access, server-side rendering, structured data (schema.org), fast load times. Get those right first. Then add llms.txt.

Who Benefits Most?

llms.txt makes the most practical difference for:

Documentation and knowledge-base sites — where AI tools are routinely used to answer user questions
SaaS and software companies — whose content is frequently cited in AI-assisted developer research
Publishers and media sites — with large archives that would otherwise be hard to summarise
Professional services firms — looking to appear accurately in AI-generated answers about their specialisms

It matters least for thin sites, sites with very dynamic content, or businesses in sectors where AI Overviews are rare.

How to Create an llms.txt File

Creating a basic llms.txt takes fifteen minutes. Here is the process:

Identify your most important pages — not all of them, just the ones that answer the questions your customers actually ask.
Write a one-paragraph summary of your site's purpose.
Group pages into logical sections (services, resources, about, etc.).
Write the file in plain Markdown following the spec.
Place it at your site root so it is accessible at yourdomain.com/llms.txt.
Validate it — the free Agent-Ready grader checks your llms.txt for spec compliance and generates a valid one for your site if you do not have one yet.

Worked Example: A Small Business llms.txt

Below is a valid llms.txt for a fictional small business — Anchor Point Electrical, a Bristol-based commercial electrician.

# Anchor Point Electrical

> Commercial electrical contractors serving Bristol and the South West since 2008. We handle installation, maintenance, and EV charging infrastructure for businesses of all sizes. NICEIC approved.

## Services

- [Commercial Electrical Installation](/services/installation): New builds, fit-outs, and rewires for commercial properties.
- [Electrical Maintenance Contracts](/services/maintenance): Planned preventive maintenance and emergency call-out cover.
- [EV Charging Installation](/services/ev-charging): Workplace EV charger supply and installation, OZEV-approved.
- [Emergency Electrician Bristol](/services/emergency): 24/7 emergency response across the Bristol area.

## About

- [About Anchor Point Electrical](/about): Company background, accreditations, and the team.
- [NICEIC Approved Contractor](/about/niceic): Our certifications and what they mean for your project.

## Resources

- [Electrical Safety Guides for Business Owners](/resources/electrical-safety): Plain-English guides to legal obligations and best practice.
- [EV Charging Grant Guide](/resources/ev-grants): Current OZEV grants available for businesses in 2025-26.

## Contact

- [Get a Quote](/contact): Request a free site survey and quotation.

Notice what is happening: every link has a description, the blockquote captures what the business does in one sentence, and the sections mirror what a prospective customer — or an AI answering "who installs EV chargers in Bristol?" — would want to find.

Checking Your AI Readiness Beyond llms.txt

llms.txt is one signal among many. The free Agent-Ready grader audits your site across the full set of AI-readiness factors — llms.txt validity, structured data coverage, crawlability, rendering method, and more — and gives you a prioritised action list.

Frequently Asked Questions

Do I need an llms.txt file?

Probably not urgently, but it is worth having. Implementation takes under an hour for most sites. Given that Google's Lighthouse now flags it and AI-agent usage is only growing, adding a valid llms.txt is cheap insurance. Start with the fundamentals (crawlability, schema, speed), then add llms.txt. If you want a file generated for you automatically, the Agent-Ready grader builds one based on your site's content.

Does ChatGPT or Google use llms.txt for rankings?

As of mid-2026, no. Neither OpenAI nor Google has confirmed that llms.txt influences what their AI products surface or how they rank content. Google has specifically said llms.txt is not required for Search. However, Google's Lighthouse tooling now checks for it under the Agentic Browsing category, and the spec is gaining traction in the developer community. Treat it as a forward-looking best practice rather than a ranking switch.

What is the difference between llms.txt and llms-full.txt?

llms.txt is a curated index — a map of your site with links and short descriptions. llms-full.txt goes further by inlining the actual content of those pages directly in the file. The full version is useful for AI tools that want to read your site's substance in one request without crawling individual URLs, but it is larger, harder to maintain, and not always necessary. Start with llms.txt. Add llms-full.txt later if you have a content-heavy site where deep comprehension matters (documentation, research, legal resources).

Jamie McKaye is an independent SEO and AI-search consultant with 19 years' experience. He implements the fixes — not just the recommendations. Get in touch if your site needs hands-on work, or run the free Agent-Ready audit to see how your site scores right now.