Short answer
llms.txt is a Markdown file at the root of a website that gives AI tools a short summary of the site and a list of its most useful pages. Jeremy Howard proposed it in September 2024. In 2026 its readers are a few AI crawlers and coding agents, Google Search ignores it, and studies find no measurable effect on AI citations. It is cheap to publish and worth keeping accurate, and it is not a ranking lever.
Key takeaways
- llms.txt is a reading guide for AI tools: a title, a short summary and lists of links with one line on each.
- It controls nothing. robots.txt decides what crawlers may fetch; llms.txt only suggests what is worth reading.
- Google Search does not use it, and Google says publishing one neither helps nor hurts.
- Ahrefs found that 97% of published files got no requests in a month, and SE Ranking found no link between the file and AI citations.
- Publish one if it costs you little, keep it accurate, and put the real effort into the content and authority that get cited.
What llms.txt is
Jeremy Howard proposed llms.txt on 3 September 2024, and the proposal lives at llmstxt.org [1]. The idea is simple. Websites are built for people and browsers, with menus, scripts and layout around the text. A language model working inside a limited context window does better with a short, clean map of what matters. llms.txt is that map, published at /llms.txt.
The only required part is an H1 with the name of the site. The proposal adds a blockquote with a short summary, optional paragraphs of context, and H2 sections that list links, each followed by an optional note after a colon. A section called Optional holds links an agent can skip when it needs to save space. The proposal also recommends a Markdown version of each important page at the same address with .md added [1].
How llms.txt differs from robots.txt and sitemap.xml
The three files sit side by side at the root of a site and do different jobs.
| robots.txt | sitemap.xml | llms.txt | |
|---|---|---|---|
| Job | Tells crawlers what they may fetch | Lists the URLs to index | Points AI tools to the most useful pages |
| Format | Plain-text rules | XML | Markdown |
| Effect | Followed by well-behaved crawlers | A hint for search engines | Optional reading for tools that choose it |
| Who reads it | Search and AI crawlers | Search engines | A few AI crawlers and coding agents |
If you want to allow or block an AI crawler, that happens in robots.txt and at your CDN or firewall. llms.txt cannot block anything.
Who reads llms.txt in 2026
Fewer readers than the attention suggests. Ahrefs checked 137,210 domains with traffic in May 2026: 28% published an llms.txt file, and 97% of those files received no requests at all that month [2]. Of the requests that did arrive, 96% came from bots, and the named AI tools fetching the file most were OpenAI’s GPTBot and Anthropic’s Claude Code [2].
Google Search is not one of them. Google’s guide to its generative AI features says you don’t need machine-readable files, AI text files or Markdown to appear in Google Search, and that creating them will neither harm nor help a site’s visibility, because Google Search ignores them [3].
Does llms.txt help you get cited by AI?
Not measurably, on the evidence so far. SE Ranking modelled AI citations for nearly 300,000 domains, 10.13% of which had an llms.txt file, and found no link between the file and how often a domain is cited. Its model became more accurate when the llms.txt variable was removed [4]. That fits the request logs: a file that is rarely fetched cannot change many answers.
Assistants cite the pages they retrieve through search and the brands the web already trusts on a topic, which I cover in what is answer engine optimization (AEO). llms.txt adds nothing to those signals. Where it does help is narrower: an agent or a developer tool that reads it gets a clean map of your key pages, which matters most on documentation and product sites.
How to create an llms.txt file
- Write a title and a summary: an H1 with the site’s name, then a one-paragraph blockquote saying who you are and what the site covers, with facts a model can quote.
- List the pages that matter in H2 sections, such as Services, Guides and Case studies, each link followed by one line saying what the page answers.
- Put secondary pages under a section called Optional.
- Publish it at the root, at /llms.txt, as plain text, and check that every link returns a 200 status and is indexable.
- Keep it in step with the site: when a page is added, renamed or removed, the file changes the same day.
A file for a small software company could look like this. The company and its pages are illustrative.
# Example Analytics
> Example Analytics is a product analytics tool for B2B SaaS teams. Founded 2019, based in Dublin.
## Product
- [Features](https://example.com/features/): What the product tracks and how teams use it.
- [Pricing](https://example.com/pricing/): Plans, limits and what each one includes.
## Guides
- [How to measure activation](https://example.com/guides/activation/): A step-by-step method with the formulas.
## Optional
- [Company news](https://example.com/news/): Announcements and release notes.On WordPress, Rank Math generates the file from your pages and posts, and this site makes its own that way. Check what it prints. On one of my sites the generated descriptions came out as a breadcrumb trail instead of a summary, so the file listed every page without saying what any of them covered. A short filter fixed it by using each page’s meta description.
What to expect from llms.txt
Expect a tidy map, not traffic. Publish it because it costs little, because some crawlers and coding agents do read it, and because the file will already be in place if more tools adopt the format. Then measure it like anything else: check your server or CDN logs for requests to /llms.txt and note which user agents fetch it. If the only visitor is your own check, the file is doing what the studies predict.
llms.txt is part of my programs, as one small technical task among many. The work that moves citations is the rest: the questions buyers ask, pages that answer them cleanly, entity and schema coverage, and mentions on the sites the engines already trust. That is where my AEO strategy and technical audit starts.
Frequently asked questions
Is llms.txt a standard?
It is a proposal, published at llmstxt.org, and no AI tool is bound to follow it. Each tool decides for itself whether to read the file.
Does Google use llms.txt?
Not for Search. Google says Google Search ignores machine-readable AI files, so they neither help nor harm a site’s visibility in Search or its AI features.
Can llms.txt block AI crawlers?
No. Crawler access is set in robots.txt and at your CDN or firewall. llms.txt only suggests what is worth reading.
Will llms.txt hurt my SEO?
No. Google Search ignores it, and other tools treat it as optional reading. The one risk is a file that lists the wrong pages or describes them badly, so keep it accurate.
Do I need llms.txt to be cited by ChatGPT?
No. The studies so far show no citation lift from the file. OpenAI’s GPTBot is among the few bots that request it, so publishing one does no harm.
Sources
- Jeremy Howard, “The /llms.txt file”, llmstxt.org, 3 September 2024 (revised 10 August 2026). llmstxt.org
- Ahrefs, “We Analyzed 137K Sites: 97% of llms.txt Files Never Get Read”, 15 June 2026. ahrefs.com
- Google Search Central, “Google’s Guide to Optimizing for Generative AI Features on Google Search”, last updated 10 July 2026. developers.google.com
- SE Ranking, “Does LLMs.txt impact your AI visibility and citations? No, according to research”, 15 October 2026. seranking.com