llms.txt Example: A Working Template for UK Sites

Last updated: 3 November 2026

A working llms.txt example is a plain-text Markdown file hosted at a site's root (yourdomain.co.uk/llms.txt) that lists a business's most important pages in a structured, AI-readable format. Adoption varies wildly by study — from under 0.005% to 57% of sites, depending on sample — and 97% of files studied received zero AI requests in one May 2026 analysis, so treat it as a low-cost addition, not a silver bullet.

Key Takeaways

Aether AI notes that an llms.txt file is a Markdown document hosted at the website root that summarises a site's content for large language models in a structured, link-based format.
A meaningful share of live websites in the Tranco top 10,000 served a valid llms.txt file, with adoption notably higher among the top 1,000 sites.
97% of llms.txt files studied across roughly 137,000 domains received zero AI requests in May 2026, according to Digitortoise (2026).
Shopify adoption of llms.txt stood out at 78.1% among CMS platforms in a top-10k crawl, per Casey Burridge's HTTP Archive analysis (2026).
Google has stated that Search and AI Overviews do not use llms.txt as a ranking or visibility signal, per analysis summarised by llmtxt.info.

97%: 97% of llms.txt files studied across roughly 137,000 domains received zero AI requests in May 2026, according to Digitortoise (2026).

What is llms.txt and what is it used for?

An llms.txt file is a plain-text Markdown document, hosted at a website's root domain, that gives large language models (LLMs) — the AI systems behind tools like ChatGPT and Claude — a concise, structured map of a site's most important content. It was proposed to solve a specific problem: LLMs have limited context windows (the amount of text they can process at once) and struggle to parse complex HTML, navigation menus, JavaScript and adverts the way a human browser does.

The specification originates from llmstxt.org, the canonical proposal page associated with Jeremy Howard. The idea borrows its name and location from robots.txt, but the function is different: robots.txt tells crawlers what they may access, while llms.txt tries to tell an AI assistant what matters most and where to find a clean, Markdown version of it.

In practice, llms.txt is mostly used by documentation-heavy businesses — SaaS platforms, API providers, developer tools — to point AI coding assistants and chat tools towards canonical product docs, pricing pages, changelogs and guides. Mintlify has documented real implementations from companies including Anthropic, Vercel, Stripe and Cloudflare, each following a similar pattern of linking out to Markdown versions of key pages.

Mintlify's own position captures why some vendors have built tooling around this: "every company will need two versions of their docs: one for humans and another for LLMs," according to a Mintlify company statement.

What does a complete llms.txt file example look like?

A complete llms.txt example is a short Markdown file starting with an H1 title, a blockquote summary, and one or more H2 sections listing links with brief descriptions. Here is a minimal but valid structure, adapted to a fictional UK software business:

# Pulse Operations

> Pulse Operations is workforce-management software for field service and facilities teams, built by the Aether Agency founding team.

Pulse Operations helps operations managers schedule shifts, track compliance and manage mobile teams from a single dashboard. The product is aimed at UK facilities, security and field-service businesses.

## Docs

- [Getting Started](https://pulseoperations.co.uk/docs/getting-started): Account setup, first schedule, and inviting team members.
- [API Reference](https://pulseoperations.co.uk/docs/api): REST endpoints for shifts, timesheets and compliance records.
- [Integrations](https://pulseoperations.co.uk/docs/integrations): Connecting payroll, HR and access-control systems.

## Pricing and Plans

- [Pricing](https://pulseoperations.co.uk/pricing): Current plan tiers and included features.

## Optional

- [Changelog](https://pulseoperations.co.uk/changelog): Recent product updates and fixes.
- [Status Page](https://pulseoperations.co.uk/status): Live uptime and incident history.

Each bullet follows a consistent pattern: a Markdown link, a colon, then a short plain-English description. The "Optional" section is a recognised convention under the specification for lower-priority links that can be skipped if a model has limited context space. Some providers, such as GitBook, also publish a companion llms-full.txt file — a single concatenated document containing the full Markdown text of every linked page, rather than just links and summaries, for models that can handle a larger context window.

What is the correct syntax and structure for formatting an llms.txt file?

The llms.txt specification requires a strict Markdown structure: one H1 title, an optional blockquote summary, optional free-text paragraphs, then H2-headed sections containing bullet-point links. The llmstxt.org specification and the more detailed AI Visibility llms.txt Specification v1.8.0 both set out the same core rules.

Required and recommended elements:

H1 title (mandatory, first line): the name of the project, product or company. Only one H1 is permitted.
Blockquote summary (recommended, directly after the H1): a single short paragraph in > blockquote format giving essential context.
Free-text body (optional): additional paragraphs or bullet lists without headings, giving background a model needs before parsing links.
H2 sections (one or more): each groups related links under a heading such as "Docs", "API", "Guides" or "Optional".
Link items: formatted as - [name](url): description, one per line, inside an H2 section.
File encoding: UTF-8 plain text, served with a .txt extension, not HTML.

A median file size of 14.2 KB was recorded across 217 notable technology, media and commerce hosts, according to llmtxt.info (2026) — useful as a rough benchmark for how much content a typical implementation includes, though file size has no bearing on validity.

Common syntax errors include skipping the H1, using H3 headings for link groups instead of H2, mixing numbered lists with the required bullet format, or linking to HTML pages instead of their Markdown equivalents where the specification calls for one.

Where should an llms.txt file be hosted on a website?

An llms.txt file must be hosted at the root of a domain, accessible at https://www.example.co.uk/llms.txt, in the same way robots.txt sits at /robots.txt. This root placement is non-negotiable under the specification: crawlers and AI agents look for the file at that fixed, predictable path rather than searching subdirectories.

For UK businesses running on common platforms, placement varies slightly by setup:

Static sites and most CMS platforms: place llms.txt directly in the public root folder alongside robots.txt and sitemap.xml.
Shopify stores: Casey Burridge's HTTP Archive analysis found Shopify adoption of llms.txt at 78.1% among CMS/platforms in a top-10k crawl, reflecting how straightforward root-level file hosting is on that platform.
Documentation platforms such as Mintlify, GitBook and Fern often auto-generate and host the file for you, sometimes alongside a split index for very large documentation sets.
Multi-subdomain businesses (e.g. a marketing site on www. and docs on docs.) may need a separate llms.txt at each subdomain root, since the file is scoped per-domain, not per-site.

Never bury the file in a subfolder like /assets/llms.txt or /files/llms.txt — it will not be found by tools looking for the standard root path.

How is llms.txt different from robots.txt and sitemap.xml?

llms.txt, robots.txt and sitemap.xml solve three distinct problems despite sitting in the same root folder. Confusing them is one of the most common mistakes UK businesses make when first adopting the format.

File	Purpose	Format	Audience
robots.txt	Tells crawlers which paths they may or may not access	Plain-text directives (User-agent, Disallow, Allow)	Search engine and AI crawlers (access control)
sitemap.xml	Lists every indexable URL on a site for discovery	XML, machine-only	Search engine crawlers (discovery)
llms.txt	Curates and summarises the most important pages for AI context	Markdown with headings, blockquotes and described links	LLMs and AI assistants (context and comprehension)

robots.txt is governed by a long-standing informal web convention and is respected by compliant crawlers as an access-control signal. Sitemap.xml, often submitted via Google Search Console, is purely about discovery — helping a crawler find every URL, with no editorial curation. llms.txt is neither an access-control file nor an exhaustive URL list; it is a hand-curated, human-readable summary intended to help a model understand what matters most on a site, written in the Markdown an LLM parses more cleanly than raw HTML.

Crucially, llms.txt is not an official web standard in the way robots.txt (formalised via RFC 9309) is. It remains a community proposal, and — per analysis summarised by llmtxt.info — Google has indicated that Search and AI Overviews do not currently use llms.txt as a ranking or visibility signal.

Which AI platforms and tools currently recognise or support llms.txt?

Support for llms.txt is inconsistent across AI platforms, and no major AI engine has publicly confirmed systematic, routine crawling of the file as of October 2026. Adoption has concentrated on the tooling side — documentation platforms that generate the file — rather than confirmed, universal consumption by AI assistants.

What is verifiable:

Google has indicated, via analysis summarised by llmtxt.info, that Search and AI Overviews/AI Mode do not use llms.txt as a ranking or visibility signal.
Documentation and developer-tool platforms — Mintlify, GitBook and Fern — all auto-generate llms.txt files for sites built on their platforms, which is one reason Technology-sector adoption is disproportionately higher than among top websites overall.
Independent consumption data is thin. The most direct evidence available is that 97% of llms.txt files studied across roughly 137,000 domains received zero AI requests in May 2026, according to Digitortoise (2026) — suggesting that whatever consumption exists today is concentrated on a small minority of sites.

For a UK business, the honest summary is: publishing llms.txt costs little, documentation platforms will often generate it for free, but there is not yet verified evidence that major consumer-facing AI assistants are systematically fetching and using it the way they use, say, structured data or a well-optimised page itself.

Is there evidence that llms.txt actually improves AI visibility or citations?

Direct evidence that publishing llms.txt improves AI citations is currently weak, and the clearest data point available argues for caution rather than confidence. The Digitortoise study found that 97% of llms.txt files studied across roughly 137,000 domains received zero AI requests in May 2026 — meaning the overwhelming majority of published files were not fetched at all during that period, let alone demonstrably influencing a citation.

Adoption figures themselves are also wildly inconsistent, which complicates any benchmarking exercise:

Published llms.txt adoption figures across studies range from under 0.005% to 57% of websites, depending on sample and methodology, according to Good For Bots (2026).
5.61% of the top 10,000 websites had a valid llms.txt as of June 2026, up from 1.04% in July 2025 — roughly 5.4x growth in 12 months, according to Casey Burridge's HTTP Archive BigQuery analysis (2026).
The HTTP Archive Web Almanac 2025 found only 2.13% of desktop home pages in a July 2025 whole-web crawl served a valid llms.txt, per HTTP Archive Web Almanac (2025).
On a curated panel of 217 notable technology, media and commerce hosts, adoption reached 57.6% as of 19 October 2026, according to llmtxt.info (2026) — a far higher figure, because that sample deliberately skews towards large, AI-savvy technology brands rather than the general web.

The gap between the 2.13% whole-web figure and the 57.6% curated-panel figure shows the real pattern: llms.txt adoption is concentrated among technology, media and SaaS companies, growing quickly in that niche, but still rare across the general web — and actual AI fetch activity appears rarer still. At Aether AI, citation tracking runs across six AI engines — ChatGPT, Perplexity, Google AI Overviews, Claude, Gemini and Copilot — precisely because a business cannot assume a technical file is working; it has to measure whether its pages are actually being named in AI answers, which is a different question entirely from whether a crawler fetched a file.

What are common mistakes businesses make when writing an llms.txt file?

The most common llms.txt mistake is treating it as a ranking lever rather than a documentation aid, which leads businesses to over-invest time relative to its proven impact. Beyond that expectation problem, several practical errors recur.

Skipping the required H1 title or blockquote summary, leaving the file structurally invalid against the llmstxt.org specification.
Linking to HTML pages where a clean Markdown version exists, forcing any model that does fetch the file to parse navigation, cookie banners and scripts anyway.
Listing hundreds of low-value links with no prioritisation, defeating the purpose of a curated summary and risking a file far beyond a sensible length.
Exposing internal, unpublished or sensitive documentation URLs — a governance risk, since anything listed becomes a discoverable pointer for both AI tools and anyone who reads the plain-text file.
Never updating the file after launch, so it goes stale as pages are renamed, merged or retired — a particular risk given the median adoption window across studies spans barely 18 months.
Assuming publication alone drives citations, when 97% of files studied received zero AI requests in May 2026, according to Digitortoise (2026).
Duplicating sitemap.xml instead of curating, listing every URL rather than the handful that genuinely represent the business's core offering.

Your llms.txt checklist

Confirm your H1 title and one-line blockquote summary are the first two elements in the file.
Group links under clear H2 sections (Docs, Pricing, Guides, Optional) rather than one flat list.
Format every link as - [name](url): description, with a genuinely useful description.
Host the finished file at yourdomain.co.uk/llms.txt, never in a subfolder.
Check the file does not expose unpublished, internal or sensitive URLs.
Cross-reference it against robots.txt and sitemap.xml so all three are consistent.
Review and update it every time you launch, rename or retire a key page.
Validate the file loads as plain text (not HTML) and check server logs periodically for fetch activity.

FAQ

What is an llms.txt example used for?

An llms.txt example demonstrates how to structure a Markdown file — H1 title, blockquote summary, then H2-headed link sections — that gives AI models a curated map of a website's key pages. It is used as a template businesses adapt to their own content rather than copied verbatim.

Is llms.txt an official web standard?

No, llms.txt is not an official, ratified web standard in the way robots.txt is. It originates from a community specification at llmstxt.org, and analysis summarised by llmtxt.info indicates Google does not currently use it for Search or AI Overviews.

Where do I host my llms.txt file?

Host llms.txt at the root of your domain, at the exact path yourdomain.co.uk/llms.txt, in the same location as robots.txt and sitemap.xml. Subfolder placements will not be found by tools expecting the standard root path.

How is llms.txt different from llms-full.txt?

llms.txt lists links with short descriptions, while llms-full.txt, as described by GitBook, concatenates the complete Markdown content of every linked page into a single file. Businesses with smaller documentation sets often publish both; larger sites sometimes publish only the indexed version to control file size.

Do ChatGPT, Claude and Perplexity actually read llms.txt files?

There is no confirmed, systematic evidence that major AI assistants routinely crawl llms.txt files. One study found 97% of llms.txt files across roughly 137,000 domains received zero AI requests in May 2026, according to Digitortoise (2026), suggesting consumption is currently limited.

How many UK and global businesses have published an llms.txt file?

Adoption estimates vary enormously depending on which sample is used, ranging from a small single-digit share of the broader web up to a notably higher share among the most prominent sites, with growth visible across more recent studies.

Should every UK business publish an llms.txt file right now?

Publishing llms.txt is low-cost and low-risk, so most businesses can reasonably add one, but it should not replace core GEO work such as structured data, clear page content and citation monitoring. Given that a meaningful share of the top 1,000 websites publish one, with Technology-category adoption notably higher, documentation-heavy and SaaS businesses have the clearest early-mover case.

Measuring what llms.txt can't tell you, with Aether AI

Publishing an llms.txt file answers one narrow question — have you given AI systems a tidy map of your site — but it says nothing about whether your business is actually being named in AI answers, which is the question most UK business owners actually care about. Aether AI's citation tracking runs across six AI engines (ChatGPT, Perplexity, Google AI Overviews, Claude, Gemini and Copilot), so a technical step like llms.txt can be checked against real citation outcomes rather than assumed to be working.

That gap between publishing a file and being cited is a live one: in Aether AI's own measured baseline of 700 live AI engine scans across UK brands' target keywords in July–August 2026, those brands were named in just 23 answers — 3.3% of scans — showing how much work sits between technical readiness and actual AI visibility, even for the founding team's own operations. Aether AI is built and operated by Aether Agency Ltd. Priority First, Pulse Operations and Aether Agency are businesses connected to the founding team.

If you want to see where your own site currently stands before deciding how much time to invest in files like llms.txt, Aether AI's free AI-visibility audit at /audit checks your current citation footprint, and the platform's public pricing covers ongoing tracking and AI-optimised content once you know where the gaps are.