llms.txt Explained With Examples

Developer outlining a site summary file for ai tools

llms.txt is a plain-text Markdown file placed at the root of a website (/llms.txt) that gives large language models a short summary of the site and a curated list of links to its most useful pages. Jeremy Howard proposed it in September 2024 as a way to help AI tools read a site without wading through navigation, ads and scripts. It is a proposal, not an adopted standard: as of October 2026 no major AI search engine has said it uses the file to choose sources, and Google says its Search does not use it.

Below you will find what the format looks like, a complete example you can adapt, what the evidence says about its effect, and a straight answer on whether your site needs one.

Where llms.txt came from

Jeremy Howard published the proposal on 3 September 2024 at llmstxt.org; the page now carries a “v2” label and was last modified in August 2026. The page links to Answer.AI and notes that Answer.AI and fast.ai software projects have published Markdown versions of their documentation under the scheme.

The reasoning is practical. Web pages are built for people, so the useful text sits inside menus, cookie banners, ads and JavaScript, and converting that back into clean text is slow and error-prone. The proposal also notes that context windows, while bigger than they were, are still too small to hold most websites whole. A short, hand-picked index in Markdown solves both problems for an AI tool that is already trying to read your site, such as a coding assistant pulling in library documentation.

That origin explains its sweet spot. The llms.txt standard was designed for developer documentation read at inference time, not as a ranking signal for AI search.

The llms.txt format, section by section

The proposal defines a small set of elements, in a fixed order. Only the first is required.

Order Element Required? What it holds
1 H1 heading (# Name) Yes The name of the site or project
2 Blockquote (> ...) No A short summary with the key facts needed to understand the rest of the file
3 Free-form Markdown No Paragraphs or lists with extra detail; any Markdown except headings
4 H2 file lists (## Section) No Links as [name](url): notes, one per line
5 ## Optional section No Secondary links an AI tool can skip when it needs a shorter context

Three conventions sit around the file itself:

  • Location. The file usually lives at /llms.txt, but it can also sit in a subpath such as /docs/llms.txt, in which case it covers the URLs under that path.
  • Markdown copies of pages. The proposal suggests offering a clean Markdown version of each important page at the same URL with .md appended (for example /pricing.html.md or /pricing.md).
  • Discovery hints. A page can point to its Markdown version with rel="alternate" type="text/markdown" and to its llms.txt file with rel="describedby", either as HTML link elements or an HTTP Link header.

You will also see llms-full.txt on some documentation sites: a single file containing the full text of the docs rather than links. It is a community convention; the llmstxt.org page does not define it.

A complete llms.txt example

Here is a full llms.txt example for a fictional invoicing software company. Copy the structure, replace the content, and keep each note to one line that tells a model what it will find behind the link.

Developer writing a plain-text markdown file for a website root
Llms. Txt explained with examples 3
# Northwind Invoicing

> Northwind Invoicing is online invoicing and time-tracking software for freelancers and agencies of up to 50 people. Founded in 2019, based in Lisbon, sold worldwide in 6 currencies. Plans start at $12 per user per month.

Key facts:
- Supports hourly, fixed-fee and retainer billing.
- Integrates with Stripe, PayPal and QuickBooks.
- Not designed for inventory or retail point-of-sale use.

## Product
- [Features overview](https://example.com/features.md): What the product does, grouped by billing, time tracking and reporting
- [Pricing](https://example.com/pricing.md): Current plans, per-user prices and what each plan includes
- [Integrations](https://example.com/integrations.md): Supported payment and accounting tools with setup notes

## Help
- [Getting started](https://example.com/docs/getting-started.md): Account setup, first client and first invoice in 10 minutes
- [Multi-currency invoicing](https://example.com/docs/currencies.md): Supported currencies and exchange-rate rules
- [API reference](https://example.com/docs/api.md): REST endpoints, authentication and rate limits

## Company
- [About](https://example.com/about.md): Team, founding story and contact details
- [Security](https://example.com/security.md): Data hosting, encryption and compliance

## Optional
- [Changelog](https://example.com/changelog.md): Release notes since 2023
- [Blog](https://example.com/blog.md): Guides on invoicing, pricing and cash flow for freelancers

A few choices in that file are deliberate. The blockquote packs the facts a model most often gets wrong about a company: category, audience, location, price and what the product is not for. Links point to .md versions so a tool gets clean text. And low-priority material sits under “Optional”, which the proposal says a tool may skip when it needs less context.

To publish the file, save it as UTF-8 plain text, upload it to the site root, and check that https://yourdomain.com/llms.txt returns a 200 status with a text content type. If you would rather not write it by hand, Asheard offers a free llms.txt generator that builds a first draft from your site’s pages, which you can then edit.

llms.txt vs robots.txt vs sitemap.xml

The three files are often confused, but they do different jobs and only one of them is honored by AI crawlers today.

Attribute robots.txt sitemap.xml llms.txt
Purpose Tells crawlers what they may or may not fetch Lists URLs for discovery and recrawl Summarizes the site and points to the best pages for AI tools
Format Plain-text directives XML Markdown
Status Long-established, widely honored standard Supported by major search engines Proposal, no formal adoption
Honored by AI search crawlers Yes, by the documented bots Used by search indexes AI engines draw on No vendor has confirmed using it for citations
Can block access Yes No No

If your goal is to control which AI bots can reach your pages, robots.txt is the file that works. Our AI crawlers list covers the token for each bot.

What the evidence says about llms.txt

The SE Ranking study of 300,000 domains

SE Ranking checked about 300,000 domains and found an llms.txt file on 10.13% of them, spread fairly evenly across traffic tiers. It then modelled how often each domain was cited by major LLMs. Simple correlation tests found no significant relationship between having the file and citation frequency, and removing the llms.txt variable from their XGBoost model made the model more accurate. The study’s conclusion was that the file does not seem to affect AI citation frequency, “at least not yet” (Search Engine Journal, November 2025). SE Ranking also observed GPTBot fetching llms.txt files only occasionally, with no link to citation outcomes.

What Google has said

  • John Mueller, June 2025: no AI system currently uses llms.txt, and server logs show chatbots fetching pages but not the file.
  • Gary Illyes: Google does not support llms.txt and has no plans to; normal SEO is enough for AI Overviews.
  • Google’s May 2026 guide to generative AI features: Google Search does not use llms.txt, though creating one for other services does no harm (Google Search Central).
  • Mueller, June 2026: called the file “purely speculative for now” and noted that because it is self-reported, it cannot help a system tell sites apart, though it might help an agent that is already on a site.

There is one mixed signal. Chrome’s Lighthouse 13.3 added an experimental “Agentic Browsing” audit that checks for an llms.txt file and calls it an emerging convention. It has no effect on rankings, and it reflects browser agents reading a site, not AI search choosing sources.

What OpenAI, Anthropic and Perplexity have said

As of October 2026, none of them has published a statement that their search products use llms.txt for ranking or citation. Their crawler documentation is about robots.txt. Practitioner log reports of OpenAI’s crawler requesting the file exist, but they are anecdotes, and a fetch is not the same as a ranking signal. If you want ChatGPT to cite you, crawl access and rankings matter far more; see how to rank in ChatGPT.

Should you add an llms.txt file?

Our verdict: add one if it costs you under an hour and you have a reason beyond AI search rankings, but do not expect more citations from it. The evidence so far shows no citation lift, Google says it ignores the file, and the other engines are silent. The cost is small, though, and there are real uses for AI tools and agents that read your site directly.

Your situation Recommendation Why
Developer docs, an API or a software library Yes, with Markdown page copies Coding assistants and agents read docs at inference time, which is what the format was built for
SaaS or B2B site with a clear product Optional, low priority A tidy summary of brand facts is cheap to keep, but expect no citation gain
Local business, blog or small ecommerce store Skip until other basics are done Crawl access, indexing and third-party mentions move AI visibility far more
Hoping it will lift AI Overviews or AI Mode No Google states Search does not use it

Whatever you decide, judge it by results. Note the date you publish the file, then compare how often AI engines mention and cite you over the following months against a fixed set of prompts. That before-and-after comparison, described in our guide to tracking brand mentions in AI assistants, is the only way to know whether the file did anything for your site.

Frequently asked questions

What is llms.txt used for?

llms.txt gives AI tools a short Markdown summary of a website and a curated list of links to its most useful pages, ideally in clean Markdown form. It was proposed to help language models read sites at inference time, for example when a coding assistant needs a library’s documentation. It does not block or allow crawlers; robots.txt still does that.

Does Google use llms.txt?

No. Google’s May 2026 guide to generative AI features says Google Search does not use llms.txt, and both John Mueller and Gary Illyes have said Google does not support it. Google adds that creating the file for other services does no harm. It has no effect on AI Overviews, AI Mode or ordinary rankings.

Does llms.txt improve AI citations?

There is no evidence that it does. SE Ranking analyzed about 300,000 domains and found no relationship between having an llms.txt file and how often a domain was cited by LLMs. Removing the variable even improved their prediction model. No AI search vendor has confirmed using the file to choose sources as of October 2026.

Where do I put the llms.txt file?

Upload it to the root of your domain so it loads at https://yourdomain.com/llms.txt, saved as UTF-8 plain text. The proposal also allows files in subpaths, such as /docs/llms.txt, covering the pages beneath that path. Check that the URL returns a 200 status and that robots.txt or your CDN does not block it.