What Is llms.txt, and Does It Help AI Visibility?

Short answer

llms.txt is a proposed Markdown file at your site root that gives language models and agents a short summary of your site and links to its most useful pages. It's read mainly by coding agents working with documentation. The evidence that it helps AI search visibility is weak: Google says its Search doesn't use it, and studies of hundreds of thousands of domains found no link to citations and almost no requests from AI retrieval bots. It's cheap to publish and fine to skip.

Key takeaways

llms.txt is a curated list of your best pages for agents. Unlike robots.txt, it controls nothing.
Google says Google Search doesn't use llms.txt, and that publishing one will neither help nor harm your visibility there.
SE Ranking found no correlation between having one and being cited, across about 300,000 domains. Ahrefs data reported by Search Engine Journal showed 97% of llms.txt files got zero requests in May 2026.
Its real audience is coding agents reading documentation. OpenAI, Anthropic and Google publish llms.txt files for their own developer docs.

What is llms.txt?

llms.txt is a proposal by Jeremy Howard, first published in September 2024 and revised as a second version in August 2026. The idea: websites are built for people, wrapped in navigation, ads and scripts, and agents with limited context windows do better with a short, clean guide. So you publish a Markdown file at /llms.txt that says what the site is and where the useful pages are. A file at the root covers the site, and a file at a subpath such as /docs/llms.txt covers the pages under it.

The format is simple. An H1 with the site's name is the only required part. After it come an optional blockquote summary and H2 sections that hold lists of links, each with an optional note. A section named Optional is, by convention, for secondary links an agent can skip when it needs a shorter context. The proposal also suggests offering clean Markdown versions of pages by adding .md to the page's URL.

# Acme Invoicing

> Invoicing software for freelancers and small studios. Plans start at $12 a month.

## Product
- [Pricing](https://acme.example/pricing): Plans, limits and what each includes
- [Acme vs Wave](https://acme.example/compare/wave): Features and prices, and where each wins

## Docs
- [API reference](https://acme.example/docs/api): Endpoints, authentication and rate limits

## Optional
- [Changelog](https://acme.example/changelog): Release notes

It isn't a replacement for anything. The proposal itself describes it as complementing a sitemap, which lists every page, and robots.txt, which sets what crawlers may access. It controls no access.

Who actually reads it?

Mostly coding agents. The proposal says llms.txt files are used most heavily for software documentation, where coding agents follow them to find API references and tutorials. The AI labs publish them for their own developer docs: we checked, and OpenAI, Anthropic, Google's Gemini API docs and Perplexity's docs all serve one today. That's a sign the format is useful for agents reading documentation. It isn't evidence that those companies' search products use your file to decide who to cite.

Does llms.txt help you get cited?

The evidence says not so far:

Source	What it found
Google's guide, updated July 2026	Google Search itself doesn't use llms.txt. Publishing one will neither harm nor help your visibility or rankings in Google Search
SE Ranking, about 300,000 domains	10.13% had a file. No correlation with how often a domain is cited by AI, and removing the variable made its prediction model more accurate
Ahrefs, reported by Search Engine Journal, 137,000 domains	Around 28% published a file, but 97% received zero requests in May 2026. AI retrieval bots made 1.1% of requests

In Ahrefs' data, the requests that did arrive came from SEO audit tools, coding agents and training crawlers far more than from the bots that fetch pages to answer questions. Ahrefs measured requests, not whether any bot acted on what it fetched, and its customers are more technical than the average site. Search Engine Journal also relays Google's John Mueller describing llms.txt as "not done for search".

Put together: there is no public evidence that llms.txt moves ChatGPT, Gemini, Perplexity, Claude or Google's AI features to cite you. The sources and pages those engines retrieve decide that.

How can you tell whether anyone reads your file?

Check your server logs for requests to /llms.txt and read the user agents. In Ahrefs' data, the requests came from SEO audit tools (21%), unidentified bots (14%), web crawlers such as Googlebot (13%) and tech-profiling tools (11%). Tools built to audit or study llms.txt files made 12% of requests, and Ahrefs saw no AI traffic on requests for /llms.txt paths that returned a 404. If you see coding agents fetching your file, it's doing its job for documentation. If you only see audit tools, it isn't doing anything for AI search.

Should you publish one?

It costs very little. If you have documentation or an API, it's reasonable, because that's where agents use it. For a marketing site, treat it as optional housekeeping, and don't let it replace the work that does move citations: pages that answer buyer questions, third-party mentions and crawler access. The free llms.txt Generator builds a valid file, and pages published with Citepoint get their own automatically.

If you publish one:

List pages that answer buyer questions: product overview, pricing, comparisons, alternatives and docs.
Write each note as one sentence on what the page answers.
Keep it accurate and current. Agents may trust what a file says, and Ahrefs found a research crawler surveying llms.txt files as a prompt-injection risk, so never auto-publish text you haven't read.
Put secondary pages under Optional.
Check that robots.txt lets the crawlers you care about in. That decides far more. See robots.txt for AI crawlers.

What should you do instead?

Spend the effort where Google's own guide and the studies point. Make sure AI search crawlers can read your site, write answer-first pages for the questions buyers ask, keep your facts consistent, and earn genuine mentions on the sites AI cites. How to get recommended by ChatGPT is a good place to start.

Frequently asked questions

Is llms.txt the same as robots.txt?

No. robots.txt tells crawlers what they may access. llms.txt is a curated guide for agents that are already allowed in. It can't block or allow anything.

Does ChatGPT, Claude or Perplexity read llms.txt?

The crawler documentation OpenAI, Anthropic and Perplexity publish doesn't mention it as an input to their search products. In Ahrefs' data, AI retrieval bots linked to ChatGPT and Perplexity made about 1% of requests for llms.txt files. Coding agents made more.

Will llms.txt hurt my SEO?

No. Google says Google Search ignores llms.txt files, so it neither helps nor harms your visibility or rankings there.

What's the difference between llms.txt and a sitemap?

A sitemap lists every indexable URL for crawlers. llms.txt is a short, curated overview with a note on each page, written to fit in a model's context.

Where do I put the file?

At the root of your domain, such as example.com/llms.txt. The proposal also allows a file at a subpath, such as /docs/llms.txt, to cover the pages under it.