llms.txt: What It Is, Examples, and Does ChatGPT Use It?

A short Markdown file that summarizes your site for AI tools. Here is how to write one in fifteen minutes, where to put it on each platform, and what the evidence says about who actually reads it.

AmIPicked · 11 min read · updated

llms.txt is a plain Markdown file at the root of your website (yourdomain.com/llms.txt) that gives AI tools a short summary of who you are and a curated list of your most important pages. You can write one in fifteen minutes. As of 17 November 2026, though, no major AI provider has said it uses llms.txt files from other websites to build answers, and Google says its AI search features ignore them.

So the honest summary is this: create the file because it is cheap and harmless, keep it accurate, and spend the rest of your time on the things that actually move AI recommendations. This guide covers both halves: how to build a good file, and what it can and cannot do for you.

What is llms.txt?

Jeremy Howard, co-founder of Answer.AI, proposed llms.txt in September 2024. His reasoning was practical. Web pages are full of navigation, cookie banners, scripts and ads, and a language model with a limited context window wastes space on all of it. A single Markdown file can hand the model the important facts and links in a format it reads easily.

The spec is published at llmstxt.org. The current version is labelled v2 and the page was last revised in August 2026, after two years of real-world use. It is still a community proposal, not a standard ratified by the IETF or W3C the way robots.txt eventually was.

llms.txt does not control access. It does not block or allow any crawler. If you want to manage which AI bots can visit your site, that belongs in robots.txt, which we cover in our guide to GPTBot and AI crawlers in robots.txt.

Does ChatGPT use llms.txt? What the evidence says in November 2026

This is the question most llms.txt guides skip or answer with optimism. Here is what is verifiable.

OpenAI documents four crawlers (GPTBot, OAI-SearchBot, ChatGPT-User and OAI-AdsBot) and how they treat robots.txt. Its documentation says nothing about reading llms.txt. OpenAI does publish an llms.txt for its own developer docs, but serving a file and consuming everyone else's are different things.

Anthropic and Google also publish llms.txt files for their developer documentation, according to the list on llmstxt.org. Neither has said its assistant reads third-party llms.txt files when answering users.

Google Search has been explicit. Its guide "Optimizing your website for generative AI features on Google Search", published in May 2026, says you do not need to create machine-readable files, AI text files, special markup or Markdown to appear in Search, AI Overviews or AI Mode. Google says it does not use them, and that creating them neither helps nor hurts.

Server logs tell the same story. In a study published on 15 June 2026, Ahrefs looked at 137,210 domains using Ahrefs Web Analytics. About 38,000 had a valid llms.txt file, and 97% of those files received zero requests in May 2026. Among AI bots, GPTBot (OpenAI's training crawler) fetched the file most often, followed by Claude Code, Anthropic's coding agent. Search and retrieval bots such as OAI-SearchBot and PerplexityBot barely touched it. Ahrefs also points out that a fetch is not proof the file was used.

Chrome's Lighthouse has an audit for llms.txt in its agentic browsing category. It only fails when your server returns an error for the file. A missing file (404) is marked "not applicable", because the file is optional.

Where llms.txt does get used is in developer tooling: coding agents, documentation platforms and setups where someone points an AI tool at a docs site. If you run a SaaS product with public docs, that is a real use case. For a local plumber or a bakery, it is a fifteen-minute housekeeping task with no proven upside for AI search.

How is an llms.txt file structured?

The spec defines a fixed order. Only the first element is required.

ElementMarkdownRequiredWhat goes in it
NameH1 heading: # NameYesYour business, product or project name
SummaryBlockquote: > textNoTwo or three sentences on what you do and for whom
DetailsParagraphs or lists (no headings)NoKey facts: location, hours, pricing basics, who you serve
File listsH2 headings, each followed by a list of linksNoImportant pages, grouped by topic
OptionalAn H2 named exactly ## OptionalNoSecondary links a tool can skip when it needs a shorter context

Each link is written as [Page name](https://url), optionally followed by a colon and a short note. The file can also live in a subfolder, such as /docs/llms.txt, in which case it describes the pages under that path.

The spec also suggests offering clean Markdown versions of useful pages at the same URL with .md added (for example /pricing.md), discoverable through a rel="alternate" type="text/markdown" link. That is worth doing for documentation sites. Most small business sites can skip it.

llms.txt example for a local business

Here is a realistic file for a fictional dental clinic. Copy it and replace the details with your own.

# Oak Street Dental

> Family and cosmetic dental clinic in Austin, Texas. We see adults and children, accept most PPO insurance plans, and offer same-day emergency appointments on weekdays.

- Address: 1200 Oak Street, Austin, TX 78701
- Hours: Monday to Friday 8am to 6pm, Saturday 9am to 1pm
- Phone: (512) 555-0100, online booking available
- New patient exam and cleaning: from $149 without insurance
- Team: 3 dentists, 1 orthodontist, Spanish spoken

## Services

- [General dentistry](https://www.example-dental.com/general/): checkups, cleanings, fillings
- [Emergency dentist](https://www.example-dental.com/emergency/): same-day slots for pain, broken teeth, lost crowns
- [Invisalign](https://www.example-dental.com/invisalign/): clear aligners, free consultation
- [Pricing and insurance](https://www.example-dental.com/pricing/): self-pay prices and accepted plans

## For patients

- [First visit](https://www.example-dental.com/first-visit/): what to bring and how long it takes
- [FAQ](https://www.example-dental.com/faq/)
- [Contact and directions](https://www.example-dental.com/contact/)

## Optional

- [Blog](https://www.example-dental.com/blog/): oral health tips
- [About the team](https://www.example-dental.com/team/)

Notice what this file does well. The summary answers the questions a patient would type into ChatGPT (where, who for, insurance, emergencies). The facts are specific and match what the website says. Marketing slogans are absent because a model cannot use "Austin's most trusted smile experts" for anything.

llms.txt example for a SaaS product

A software company usually has more pages, and documentation is where llms.txt sees most real use. Group links by what a user or developer is trying to do.

# Ledgerly

> Ledgerly is invoicing and expense software for freelancers and agencies of up to 50 people. It connects to Stripe and QuickBooks and has a REST API.

Plans start at $12 per user per month. A 14-day free trial needs no credit card.

## Product

- [Features](https://www.example-saas.com/features/): invoicing, recurring billing, expense tracking, time tracking
- [Pricing](https://www.example-saas.com/pricing/): plans and limits
- [Integrations](https://www.example-saas.com/integrations/): Stripe, QuickBooks, Xero, Zapier

## Docs

- [Quickstart](https://docs.example-saas.com/quickstart.md): create an account and send the first invoice
- [API reference](https://docs.example-saas.com/api.md): authentication, endpoints, rate limits
- [Webhooks](https://docs.example-saas.com/webhooks.md): events and payload formats

## Policies

- [Security](https://www.example-saas.com/security/): SOC 2 report, data residency
- [Changelog](https://www.example-saas.com/changelog/)

## Optional

- [Blog](https://www.example-saas.com/blog/)
- [Comparison with competitors](https://www.example-saas.com/compare/)

The docs links point to .md versions, which is what the spec recommends when you have them. If your docs platform does not produce Markdown pages, link the normal HTML pages instead.

What should you put in llms.txt, and what should you leave out?

Include facts a customer checks before buying and that an AI tool could quote:

  • what you do, in one plain sentence using your customers' words;
  • location and service area, if you are a local business;
  • starting prices, if they are public on your site;
  • links to service or product pages, pricing, FAQ and contact.

Leave out slogans, legal pages, login and cart pages, tag archives and anything that is not on your website. Contradicting your own site is worse than saying nothing. Treat the file as public too: it is readable by anyone, so never put internal notes, staging URLs or private pricing in it.

The last rule is upkeep. A file with last year's prices or an old address misleads any tool that reads it. Add it to whatever checklist you use when you change prices or move.

What is llms-full.txt?

llms-full.txt is a companion file that contains the full text of your important pages in one long Markdown document, so a tool can load everything in a single request. It is common on documentation sites; Mintlify and GitBook generate it automatically.

The llmstxt.org spec does not define llms-full.txt. It grew up as a convention among docs platforms. For a service business it adds maintenance and duplicates your website, so skip it unless you publish technical documentation.

How to create llms.txt (step by step)

  1. List your 5 to 15 most useful pages. Think of the questions customers ask before buying and pick the pages that answer them.
  2. Write the H1 and the summary. Two or three sentences, factual, no adjectives you could not prove.
  3. Add key facts as a short list: location, hours, pricing basics, who you serve.
  4. Group the links under H2 headings and add a short note after each one.
  5. Put secondary pages under ## Optional.
  6. Save as UTF-8 plain text named llms.txt. Avoid word processors, which can add formatting.
  7. Upload it to your site root and test it (see below).

If you would rather not write it by hand, our free llms.txt generator reads your sitemap, suggests a name, summary and page list, and lets you choose what goes in before you download the file.

How to add llms.txt in WordPress

With an SEO plugin. Yoast SEO and All in One SEO both generate llms.txt automatically, and both are listed as integrations on llmstxt.org. Look for the llms.txt option in the plugin's settings; if you cannot find it, update the plugin and search the settings for "llms.txt". After turning it on, open yourdomain.com/llms.txt and read the result. Automatic files usually pull the summary from your homepage meta description, which may be too vague.

Manually. Create the file yourself and upload it by SFTP or your host's file manager to the WordPress root folder, the one that contains wp-config.php (often public_html). If an SEO plugin also generates llms.txt, switch that feature off so you do not end up with two competing versions.

How to add llms.txt in Shopify

Shopify now serves AI discovery files itself. According to Shopify's developer documentation, stores have a platform-generated /agents.md, and /llms.txt and /llms-full.txt mirror it unless you override them. So before you do anything, open yourstore.com/llms.txt and read what is already there.

You have two ways to change it:

  • Theme templates. Shopify supports templates/agents.md.liquid, templates/llms.txt.liquid and templates/llms-full.txt.liquid. For /llms.txt, Shopify checks llms.txt.liquid first, then agents.md.liquid, then its generated default. Add llms.txt.liquid only if you want /llms.txt to differ from /agents.md.
  • An app. Yoast SEO for Shopify added an llms.txt feature in March 2026 (Settings, then Advanced, then llms.txt), with an automatic mode built from your best-selling products, largest collections and policies, or a manual page selection.

Older guides recommend uploading the file to Shopify Files and redirecting /llms.txt to it. With native support in place, that approach is outdated.

How to add llms.txt in Next.js and other frameworks

For a static file, put llms.txt in the public folder. Next.js serves everything in public from the site root, so it will be live at /llms.txt after your next deploy. The same works in Astro, Vite and most static site generators.

If you want the file built from your CMS or product data, use a route handler instead, for example app/llms.txt/route.ts in the App Router, returning a Response with Content-Type: text/plain; charset=utf-8. Do not keep both a static file and a route for the same path.

On Wix, llmstxt.org lists Wix as having built-in support, so check the SEO settings before uploading anything. On other hosted builders, search the help center for "llms.txt", since root-level uploads are often restricted.

How to test your llms.txt file

  1. Open it in a browser. yourdomain.com/llms.txt should show plain text with readable characters. If you see your homepage, the file is not at the root and your server is redirecting unknown URLs.
  2. Check the status code and content type. Run curl -I https://yourdomain.com/llms.txt. You want 200 and a text type such as text/plain or text/markdown, not text/html.
  3. Fix broken characters. Garbled accented letters or curly quotes mean the file was not saved as UTF-8.
  4. Click every link. Links that redirect or return 404 make the file less useful and look neglected.
  5. Run Lighthouse in Chrome DevTools. Its llms.txt audit flags server errors when fetching the file.
  6. Check robots.txt at the same time. A perfect llms.txt does nothing if your robots.txt or firewall blocks the AI crawlers that would visit your pages. Our free AI crawler checker shows both in one pass.

Common llms.txt mistakes

  • Saving the file from a word processor, so it contains formatting or the wrong encoding.
  • Two versions live at once: one from a plugin, one uploaded by hand.
  • A summary copied from an ad slogan that never says what the business does.
  • Listing 200 links. The point is curation; that is what your sitemap is for.
  • Never updating it after prices, hours or the address change.

llms.txt vs robots.txt vs sitemap.xml

FileAudienceJob
robots.txtAll crawlers, including AI botsSays which parts of the site a crawler may visit
sitemap.xmlSearch enginesLists every URL you want indexed
llms.txtLanguage models and AI agentsSummarizes the site and points to a handful of key pages

None replaces the others. If robots.txt blocks OAI-SearchBot, for example, ChatGPT search cannot cite your pages however good your llms.txt is.

What actually gets you mentioned by AI

If you have an hour to spend on AI visibility, give llms.txt fifteen minutes and use the rest on things with evidence behind them: crawlable service pages with clear prices and locations, reviews on the platforms AI tools cite, and mentions on directories and articles in your niche. Our guide on how to rank in ChatGPT walks through that plan, and the generative engine optimization guide explains the wider picture.

To see where you stand before changing anything, run the free AmIPicked AI visibility test. It writes five questions your customers might ask, puts them to ChatGPT and Gemini with live web search, and shows how often you are mentioned and who appears instead. The same report checks your robots.txt, llms.txt and structured data, so you can confirm the file you just created is reachable.

Frequently asked questions

What is llms.txt?

+

llms.txt is a plain Markdown file served at yourdomain.com/llms.txt. It gives language models a short summary of your site and a curated list of your most useful pages with one-line descriptions. Jeremy Howard of Answer.AI proposed the format in September 2024, and the spec lives at llmstxt.org.

Does ChatGPT use llms.txt?

+

OpenAI has not said that ChatGPT or its crawlers use llms.txt files from other websites when answering questions. In Ahrefs' June 2026 study of roughly 38,000 sites with a valid file, 97% received no requests for it in May 2026. OpenAI publishes an llms.txt for its own developer docs, which is a different thing from reading yours.

Does Google use llms.txt for AI Overviews?

+

No. Google's guide to generative AI features in Search, published in May 2026, says you do not need new machine-readable files, AI text files or Markdown to appear in Google Search, including its AI features. Google adds that such files neither help nor hurt.

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

+

llms.txt is an index: a summary plus links. llms-full.txt is a single large file containing the full text of the important pages, so a tool can load everything in one request. The llmstxt.org spec does not define llms-full.txt; documentation platforms such as Mintlify and GitBook popularized it.

Is llms.txt worth creating for a small business?

+

It takes about fifteen minutes and cannot hurt, so it is a reasonable tidy-up task. Do not expect it to change whether ChatGPT or Gemini recommend you. Reviews, clear service pages and mentions on other sites have a much stronger effect on AI answers.

How do I check if my llms.txt works?

+

Open yourdomain.com/llms.txt in a browser and confirm you see plain text, not your homepage or a 404 page. Then check that the server returns status 200 and a text content type. A free crawler checker can confirm the file is reachable alongside your robots.txt rules.

Read next