free toolno sign-uprespects robots.txt

Free llms.txt generator.

Type your site’s address and get a draft llms.txt, built from your own pages, to review and publish.

We read the home page, the sitemap and up to 30 pages. It takes up to 20 seconds.

short answer · updated 19 October 2026

An llms.txt is a markdown text file at the root of a site (yourbrand.com/llms.txt) that summarises the site for language models and AI agents: the name, a sentence on what the company does and lists of links to the pages that matter. It is an open proposal, published at llmstxt.org, not a standard. Today, no major answer engine officially confirms that it uses llms.txt, and Google says Google Search ignores it. It is cheap to add and does no harm, but it does not replace what the engines say counts: indexed pages, text in the HTML, structured data that matches the visible text and a company described the same way everywhere. This generator reads your site as a polite reader, respects robots.txt and builds the draft in seconds; reviewing and publishing it is up to you.

the format

One H1, one blockquote and lists of links.

In the order of the specification, version 2 (August 2026), read on 19 October 2026.

H1 with the name: The only required part: # Site or company name.
Blockquote: One line starting with >, with a short summary and what is needed to understand the rest.
Text, no headings: Paragraphs or lists with more detail and how to read the links. No headings in this block.
H2 sections: Each with a list of links: - [name](url): notes. Notes are optional.
Optional: By convention, the section with secondary links that an agent can skip when it is short of space.

# Example Store

> Handmade shoes from northern Portugal, since 1985.

## Main pages

- [About us](https://yourbrand.com/about/): Three generations of shoemakers.
- [Contact](https://yourbrand.com/contact/): Address, opening hours and support.

## Products

- [Alfama boot](https://yourbrand.com/products/alfama-boot/): Natural leather, stitched sole.

## Optional

- [Privacy policy](https://yourbrand.com/privacy/)

how the generator reads your site

Like a polite reader, and only what is public.

robots.txt firstWe read robots.txt and obey it for every address and every redirect. Our reader is called SO-MA-llms-txt; with no rules for it, the * rules apply. If robots.txt does not answer, we read nothing.
Home pageSite name (structured data, og:site_name or title), description, language and versions in other languages.
SitemapThe one in robots.txt or /sitemap.xml, up to 6 files. With no sitemap, the links on the home page.
Up to 30 pagesPicked to cover each part of the site (pages, services, collections, products, articles, help), those linked from the home page first. From each, only the title and the description.
Your current llms.txtIf there is one, we validate its shape against the specification and list the pages it is missing.
LimitsPublic pages only, 20 seconds in total, pages read 5 at a time (one by one, with the requested pause, if robots.txt sets a Crawl-delay), and nothing stored: not the address, not the pages, not the draft.

after the draft

Where to put it and how to link it, in 3 steps.

No promises: publishing the file does not guarantee any engine reads it. It makes sure that, if one does, it finds what the company wants known.

01

Put the file at the root of the domain

Publish the text at yourbrand.com/llms.txt as plain text (text/plain or text/markdown, UTF-8), answering 200 without redirecting to an HTML page. The specification also allows an llms.txt in a subfolder, which only covers the pages under that folder; for the whole company, use the root.
02
Link it from your pages

Version 2 of the specification recommends pointing to the file with rel="describedby": in the <head> of each page, <link rel="describedby" href="/llms.txt">, or in an HTTP header, Link: </llms.txt>; rel="describedby". rel="alternate" type="text/markdown" is for something else: pointing to the markdown version of the page itself (page.md), if you have one.
```
<link rel="describedby" href="/llms.txt">

Link: </llms.txt>; rel="describedby"
```
03
Let the robots reach it

Check that robots.txt does not disallow /llms.txt for the readers you want to let in. There is no official robots.txt line to announce llms.txt: the standard (RFC 9309) defines user-agent, allow and disallow, and the llms.txt specification adds none. A comment line with the address (starting with #) is ignored by robots but helps people. Then open the address in a browser and run it through this generator again: it reads and validates the file you already have.
```
User-agent: *
Allow: /

# Site summary for language models: https://yourbrand.com/llms.txt
Sitemap: https://yourbrand.com/sitemap.xml
```

by platform

Shopify, Wix, WordPress or a custom site.

What each platform’s documentation says, read on 19 October 2026.

Shopify: Every store already serves an /llms.txt, a copy of /agents.md with instructions for shopping agents, and Shopify recommends leaving it as is in most cases. To make /llms.txt say something else, create the templates/llms.txt.liquid template in the theme (Online Store > Themes > Edit code); only /llms.txt changes and /agents.md stays Shopify’s. Whoever creates it is responsible for keeping it current.Shopify, llms.txt.liquid (theme documentation) · Shopify developer changelog, Customize /llms.txt, /llms-full.txt and /agents.md (28 May 2026)
Wix: Wix generates and maintains an llms.txt for every site with a paid plan, a custom domain and indexing turned on. You edit it in the dashboard, under SEO & GEO; once edited it stops updating by itself, and you can reset it to the automatic version.Wix Help Center, Understanding Your Site’s llms.txt File
WordPress: The Yoast SEO and AIOSEO plugins have an llms.txt generator. Without a plugin, upload the file to the root folder of the hosting so it opens at yourbrand.com/llms.txt.Yoast SEO, llms.txt · AIOSEO, LLMs.txt Generator
Custom-built site: On a static or hand-coded site, the file goes in the public folder (public/ in Astro, Next.js or Vite) or is generated at build time from the same sources as the pages, which is what SO-MA does on its own site.

who reads llms.txt today

No major answer engine confirms it reads it.

Only what the official pages say, read on 19 October 2026. Where the source is silent, so are we.

Google Search: Google says you do not need to create new machine-readable files, AI text files, markup or Markdown to appear in Google Search, including its generative AI features, because Search does not use them. It adds that keeping an llms.txt for other services neither harms nor helps visibility in Search, because Search ignores it.Google Search Central, Google’s Guide to Optimizing for Generative AI Features on Google Search (updated 10 Jul 2026)
Chrome Lighthouse: Chrome’s audit tool has an llms.txt check in its agentic browsing group. It only flags an error if the server fails to return the file; if there is none (404), the check is marked not applicable because, the page says, providing the file is optional at the moment.Chrome for Developers, Lighthouse, llms.txt audit (updated 5 May 2026)
OpenAI, Anthropic, Microsoft and Perplexity: Each company’s official pages about its robots and search, read on 19 October 2026, talk about robots.txt and do not say they read llms.txt. None confirms using it to choose or cite sources.OpenAI, Overview of OpenAI Crawlers · Claude Help Center, Does Anthropic crawl data from the web, and how can site owners block the crawler? (7 Apr 2026) · Bing Webmaster Guidelines · Perplexity, Perplexity Crawlers
Who publishes one: OpenAI, Anthropic and Google publish an llms.txt for their own developer documentation. Publishing is not reading: it says nothing about what ChatGPT, Claude or Gemini use when they answer.OpenAI documentation llms.txt · Anthropic documentation llms.txt · Gemini API documentation llms.txt
The proposal’s author: In version 2 (August 2026), the author of the proposal writes that llms.txt is used most heavily for software documentation, by coding agents that follow it to find references, and that the same structure works for any site where an agent needs a path into the content. That is the view of whoever proposed it, not of an engine.llmstxt.org, The /llms.txt file, v2 · llmstxt.org, Changes (v2, August 2026)

In short: Today, no major answer engine officially confirms that it uses llms.txt, and Google says Google Search ignores it. It is cheap to add and does no harm, but it does not replace what the engines say counts: indexed pages, text in the HTML, structured data that matches the visible text and a company described the same way everywhere.

what the draft does not know

Five things only the site owner can confirm.

What the company does, in its own words: the blockquote at the top is the home page description, written for Google and not for an assistant.
Which pages matter most to the business. We picked up to 30 from the site’s structure; swap, merge or cut.
Facts only the company knows and an assistant should get right: areas served, lead times, terms, contacts. If they are public, add them as text before the first section.
Whether the site has markdown versions of its pages. The specification recommends linking to them (page.md); the draft links to the HTML pages.
Whether the descriptions are right and current. They come from the pages themselves, as they are today.

ai visibility · showing up in ai answers

Want us to do this for you, with the entity and the structured data?

llms.txt is the smallest part. What the engines say counts is the rest: pages indexed in Google and Bing, text in the HTML, organisation data that matches the visible text and a company described the same way everywhere.

We start with the audit, asking ChatGPT, Gemini, Perplexity and Google the questions your customers ask. Every project gets a written proposal with scope, timeline and a fixed price.

Start with the audit See the AI visibility audit

robots.txt with the search crawlers allowed on purpose
An llms.txt with what the company does, reviewed by hand
Organisation data and official profiles linked, so the company is not confused with another of a similar name
Pages that answer customers’ questions, with the FAQs in visible text
For a store: the free agent-ready checker

frequently asked

Questions about llms.txt.

Answered in full, in visible text.

What is llms.txt?

llms.txt is a markdown text file at the root of a site (for example yourbrand.com/llms.txt) that summarises the site for language models and AI agents: a heading with the name, a sentence on what the company does and lists of links to the pages that matter, each with a short note. It is an open proposal, published at llmstxt.org in September 2024 and revised in August 2026 (version 2), not a standard. It does not replace robots.txt, which tells robots what they may read, or the sitemap, which lists every page.

Do ChatGPT and Google read llms.txt?

Neither confirms it. Google says, in its official guide to the generative AI features of Search (updated 10 July 2026), that Google Search does not use files like llms.txt and that keeping one neither helps nor harms visibility in Search. OpenAI, Anthropic, Microsoft and Perplexity do not say, on their official pages about their robots and search read on 19 October 2026, that they read it. According to the proposal’s author, coding agents reading software documentation use it most. It is cheap and harmless to publish, but it is not what puts a company in the answers.

Where does llms.txt go, and how do I link it?

At the root of the domain, as plain text. Version 2 of the specification recommends linking it from each page with <link rel="describedby" href="/llms.txt"> in the HTML head, or with an HTTP Link header with rel="describedby". There is no official robots.txt line for it: robots.txt just must not disallow the address. On Shopify, /llms.txt already exists (a copy of /agents.md, for shopping agents) and is changed with the templates/llms.txt.liquid theme template.

Does the generator store my site or the draft?

No. The function reads the public pages at that moment, builds the draft and returns it to your browser; no copy of the address, the pages or the text is kept. The IP address of whoever uses it is only kept in memory, for a few minutes, to limit abuse. If you ask Lucas for help, the site address goes in your request, and only then.

Why does the generator say my site does not allow reading?

Because the site’s robots.txt disallows automated readers on the home page, or the firewall refuses robots, or robots.txt did not answer. In every case we read nothing: that is the robots.txt rule (RFC 9309), the same one we ask others to follow on our site. To get the draft, let the SO-MA-llms-txt reader in through robots.txt.

Free llms.txt generator.

Want us to do this for you, with the entity and the structured data?

One H1, one blockquote and lists of links.

Like a polite reader, and only what is public.

Where to put it and how to link it, in 3 steps.

Put the file at the root of the domain

Link it from your pages

Let the robots reach it

Shopify, Wix, WordPress or a custom site.

No major answer engine confirms it reads it.

Five things only the site owner can confirm.

Want us to do this for you, with the entity and the structured data?

Questions about llms.txt.

What is llms.txt?

Do ChatGPT and Google read llms.txt?

Where does llms.txt go, and how do I link it?

Does the generator store my site or the draft?

Why does the generator say my site does not allow reading?