Last updated 5 Nov 2026 · reviewed by the Essentras editorial team

Free tool, no sign-up

Free llms.txt generator and checker

The short answer

llms.txt is a proposed markdown file at /llms.txt that gives AI agents a short map of your site: an H1 with the name, a 1-line summary, then H2 sections of links. This free tool builds or checks one in your browser. Google says Google Search ignores it, so expect no ranking change.

Checked on 4 Nov 2026 against llmstxt.org (version 2), Google’s and Chrome’s documentation, and the crawler pages of OpenAI, Anthropic and PerplexityWritten by the Essentras editorial team

Essentras sells an AI Visibility Tracker ($49 a month) and AI SEO work, and is not affiliated with Answer.AI, Google, OpenAI, Anthropic or Perplexity. The generator is free and needs no account.

Build your file Who reads llms.txt?

Generator and checker

Generate or check an llms.txt file

Fill in the form or paste your URLs, then copy or download the file. To audit a file you already have, paste it into the checker. Every rule the tool applies is explained below the form.

Runs in your browser. Nothing you type is sent to Essentras or to any server, and the tool never visits your URLs.

Site or project name

Becomes the H1, the only line the format requires.

1-line summary

Becomes the blockquote: what you are, plus the facts an agent needs to read the rest.

Details (optional)

Plain paragraphs or lists. No headings: a heading here would break the format.

Sections of links

Each section becomes an H2 with 1 line per link. Point to the pages that answer real questions, ideally in clean markdown.

Optional section

Secondary links an agent can skip when it needs a shorter context. Written last, under "## Optional".

Your llms.txt

Add a name and a 1-line summary to see your file.

Next: see what AI assistants say about you

llms.txt is a map, not a ranking signal. To see whether assistants name you, ask them: the free AI check runs 3 buyer questions live, and the AI Visibility Tracker asks 7 assistants every week.

Run a free AI check AI Visibility Tracker: free, or $49 a month

Key facts, checked on 4 Nov 2026

llms.txt is a proposal, not a web standard. Jeremy Howard of Answer.AI published it on 3 Sep 2024; version 2 is dated 10 Aug 2026.
llmstxt.org
The H1 with the site or project name is the only required part. The summary, the details and the H2 sections of links are optional.
llmstxt.org, Format
Google says Google Search ignores llms.txt files: keeping one neither helps nor harms rankings. Google added that note on 15 Jun 2026.
Google Search Central
Chrome’s Lighthouse has an experimental llms.txt audit: a server error is flagged, and a missing file is marked not applicable because the file is optional.
Chrome for Developers
OpenAI, Anthropic and Perplexity publish llms.txt files for their own docs, but their crawler pages do not say their bots read yours.
Who reads it, table below
In Ahrefs’ data on 137,210 sites, 97% of about 38,000 llms.txt files got no request at all in May 2026.
Ahrefs, 15 Jun 2026
On 4 Nov 2026, 327 of the 3,840 media sites in the Essentras catalog (8.5%) served an llms.txt that starts with an H1; 2,249 had none.
Our survey
73 of the files we read put a “Generated by” plugin line above the H1, and 27 held robots.txt rules instead of the format.
Our survey
llms-full.txt is not part of the proposal. It is a convention from documentation platforms: the full text of every page in 1 file.
Mintlify docs
llms.txt controls no crawler. Each AI company documents its own robots.txt tokens, such as GPTBot, ClaudeBot and PerplexityBot.
AI crawler tokens, table below

The format

What llms.txt is, per llmstxt.org

What the proposal says, read on 4 Nov 2026. Version 2 is dated 10 Aug 2026.

llms.txt is a markdown file that gives AI agents a curated map of a website: who you are in 1 line, then lists of links to the pages that matter. Jeremy Howard of Answer.AI proposed it on 3 Sep 2024.

The reasoning: web pages wrap their content in navigation, scripts and ads, while agents work best with concise text in 1 place. So the file stays small, and the detail lives behind its links, which an agent fetches only when it needs them. The proposal asks for links to agent-friendly content, such as markdown copies of pages.

The file sits at /llms.txt, or under a path such as /docs/llms.txt, where it covers only the pages below that path. The proposal says it is used most for software documentation, and that the same structure suits a business, a personal site or a school.

It is not robots.txt: robots.txt tells crawlers what they may fetch, while llms.txt grants and blocks nothing. It is not a sitemap either: a sitemap lists every page, and llms.txt picks the few an agent needs.

The format, part by part

Part	Required?	How to write it
Byte-order mark	No	Allowed at the very start; most files have none.
H1 with the site or project name	Yes, the only required part	1 line: # Name
Blockquote summary	No	1 line: > what you are, and the facts needed to read the rest
Details	No	Paragraphs or lists, any markdown except headings
H2 sections (file lists)	No	## Name, then 1 list item per link: - [title](url): optional note
Optional	No	An H2 named Optional for links an agent can skip. Since version 2 it is only a convention.

Source: llmstxt.org, Format section, read 4 Nov 2026.

What changed in version 2 (August 2026)

Discovery: rel="alternate" type="text/markdown" points to a page’s markdown copy, and rel="describedby" points to the llms.txt that covers it, as HTML link tags or an HTTP Link header.
Markdown copies may add .md to the URL (page.html.md) or replace the extension (page.md).
A file covers the pages under its path, and the most specific file applies.
The tool that expanded a file into a context file is no longer part of the proposal, so the Optional section is now only a convention for secondary links.

Copy-ready llms.txt template

When to use it: Replace the example values, keep 1 link per line, and save the file as llms.txt in UTF-8.

llms.txt

# Example Co

> Example Co sells handmade ceramics online and ships to 30 countries. Prices are in USD.

Orders ship in 2 to 5 business days. Last reviewed: November 2026.

## Products

- [Catalog](https://example.com/catalog.md): every product with its price and stock
- [Shipping and returns](https://example.com/shipping.md): rates, delivery times and the returns policy

## Company

- [About](https://example.com/about.md): who we are and how to contact us

## Optional

- [Blog](https://example.com/blog.md): care guides and studio news

Example values only. The links point to markdown pages, as the proposal suggests; plain HTML pages work too.

Who reads it

Who reads llms.txt, and who says nothing about it

What each company documents, read on 4 Nov 2026. Where a company says nothing, the table says so.

Who	What they document about llms.txt	Date on the page	Source
Google Search, including AI Overviews and AI Mode	Ignores llms.txt files. Keeping one neither helps nor harms visibility or rankings, and it is fine to keep one for other systems.	Note added 15 Jun 2026; page updated 10 Jul 2026	AI optimization guide
Chrome Lighthouse (Google)	An experimental Agentic Browsing audit requests /llms.txt. A server error is flagged; a missing file is marked not applicable because the file is optional. The category needs Chrome 150 or later.	Updated 5 May 2026	Lighthouse llms.txt audit
OpenAI: OAI-SearchBot, GPTBot, ChatGPT-User, OAI-AdsBot	Its crawler page covers robots.txt controls only. It says nothing about reading a site’s llms.txt. OpenAI publishes llms.txt files for its own developer docs.	No date shown	Overview of OpenAI crawlers
Anthropic: ClaudeBot, Claude-SearchBot, Claude-User	Its crawler article covers robots.txt and Crawl-delay, and says nothing about llms.txt. Anthropic publishes llms.txt and llms-full.txt for its developer docs.	Updated 7 Apr 2026	Anthropic help center
Perplexity: PerplexityBot, Perplexity-User	Its crawler page covers robots.txt and firewall settings, and says nothing about reading a site’s llms.txt. Perplexity publishes llms.txt for its own docs.	Updated 29 Oct 2026	Perplexity crawlers
The proposal’s author (Answer.AI)	Says thousands of sites publish the file, coding agents follow it to find API references, and documentation platforms generate it.	Version 2, 10 Aug 2026	llmstxt.org

What Google says, in its own words

Google’s guide to optimizing for generative AI features lists llms.txt under the things you can ignore for Google Search. It says that keeping a file will neither harm nor help your visibility or rankings, “as Google Search ignores them”, and that it is fine to keep one for other services that use it. Google added the note on 15 Jun 2026; the page was last updated on 10 Jul 2026 and read on 4 Nov 2026.

Read Google’s note

What server logs show: Ahrefs, May 2026

Ahrefs read the server logs of the sites in its web analytics product (published 15 Jun 2026, updated 27 Aug 2026). Its sample leans technical, so Ahrefs calls the adoption figure an upper bound.

What Ahrefs measured	Result
Sites studied	137,210
Sites that publish an llms.txt	28%, about 38,000
Files that got no request at all in May 2026	97%
Requests made by bots rather than people	96%
Requests from SEO audit tools	21.7%
Requests from AI training crawlers (GPTBot, ClaudeBot)	5.3%
Requests from AI assistants fetching for a user (ChatGPT-User, Claude-User)	2.5%
Requests from AI search and retrieval bots (OAI-SearchBot, PerplexityBot)	1.1%
AI bot requests for an llms.txt that did not exist	None

Source: Ahrefs, “We Analyzed 137K Sites: 97% of llms.txt Files Never Get Read”, read 4 Nov 2026. Shares of requests are of all requests to llms.txt files that got any.

llms.txt controls no crawler: robots.txt does

To decide which AI crawlers may read your site, use the robots.txt tokens each company documents. What blocking each token does, in the company’s own terms:

Token	Company	What blocking it does, per the company
GPTBot	OpenAI	Signals that your content should not be used to train its foundation models.
OAI-SearchBot	OpenAI	Your pages are not shown in ChatGPT search answers, though they can still appear as navigational links.
ChatGPT-User	OpenAI	Visits made for a user’s request. OpenAI says robots.txt rules may not apply to them.
ClaudeBot	Anthropic	Signals that your future pages should be excluded from training data.
Claude-SearchBot	Anthropic	Your pages are not indexed for Claude’s search, which may reduce your visibility there.
Claude-User	Anthropic	Claude cannot fetch your pages when a user asks it to.
PerplexityBot	Perplexity	Your pages may not appear in Perplexity’s search results. Perplexity says this bot does not collect training data.
Perplexity-User	Perplexity	Visits made for a user’s question. Perplexity says it generally ignores robots.txt.
Google-Extended	Google	Your content is not used to train future Gemini models or to ground answers in Gemini Apps and Vertex AI. No effect on Google Search.

Sources: the crawler pages of OpenAI, Anthropic and Perplexity, and Google’s common crawlers page (updated 30 Oct 2026), all read 4 Nov 2026.

robots.txt lines for AI crawlers

When to use it: An example of 1 possible choice: AI search crawlers in, training crawlers out. Keep only the lines that match your own decision.

robots.txt

# AI search crawlers: allowed
User-agent: OAI-SearchBot
Allow: /

User-agent: Claude-SearchBot
Allow: /

User-agent: PerplexityBot
Allow: /

# AI training crawlers: not allowed
User-agent: GPTBot
Disallow: /

User-agent: ClaudeBot
Disallow: /

User-agent: Google-Extended
Disallow: /

Blocking Google-Extended also stops grounding in Gemini Apps, per Google. Blocking a search crawler can remove you from that assistant’s search answers.

Our data

llms.txt on 3,840 media sites: our survey

On 4 Nov 2026 we requested /llms.txt once from every site in the Essentras catalog of media outlets (3,888 listings on 3,840 sites) and read the first 4 KB of each answer.

What /llms.txt returned

Answer	Sites	Share of 3,840
A text file that starts with an H1, the only required part	327	8.5%
A text file with something else on top, or empty	119	3.1%
An HTML page with status 200, often the home page	214	5.6%
A redirect to another page	2	0.1%
No file: status 404 or 410	2,249	58.6%
Refused our identified survey bot, or failed	929	24.2%

Refused or failed: 502 answered 401, 403, 429 or a 5xx error, 225 answered 402 (Payment Required), 166 had a network error or timed out, 36 gave another status.

By Domain Authority (DA, as listed in our catalog)

DA band	Sites	Starts with an H1	Share of all	Share of those that answered
80 and above	747	59	7.9%	9.9%
60 to 79	2,333	189	8.1%	10.7%
40 to 59	518	57	11.0%	16.1%
under 40	242	22	9.1%	11.7%

Higher authority did not mean more adoption: the largest sites were no more likely to publish a file.

Inside the 327 files that start with an H1

Measure	Files
A blockquote summary in the first 4 KB	260
At least 1 "- [title](url)" link in the first 4 KB	288
Both	237
Neither (for example, AI usage policies written as prose)	16
Median size (201 files that stated a size)	3,243 bytes (3.2 KB)
Under 2 KB	61
Over 100 KB	18
Over 1 MB	4 (the largest: 21.7 MB)
Served as text/plain	323
Served as text/markdown	4

The 119 text files that do not start with an H1

What sat at the top	Files
A “Generated by” line from a WordPress SEO plugin, above the H1 (Rank Math SEO 45, All in One SEO 17, Yoast SEO 11)	73
robots.txt rules (User-agent, Disallow) instead of the format	27
Nothing: an empty file	9
Other: an HTML tag or comment above the H1, a code fence, a bare list of links, a blockquote with no H1, server source code, 1 unreadable compressed file	10

What the survey shows

Adoption is low: about 1 media site in 12 serves a file that starts the way the format asks (8.5%).
Many answers are not what the format describes: 333 sites return an HTML page or a text file that does not start with an H1, about as many as the 327 that do.
Among the text files, the most common defect was a line that WordPress SEO plugins add above the H1.
Most files are small: the median is about 3.2 KB, in line with the proposal’s point that the file should stay short.

Limits: 1 request per site between 18:22 and 18:24 UTC on 4 Nov 2026, with a user agent that names our survey. We did not retry the 929 sites that refused it or failed, so their files are unknown, and a site may answer browsers differently. These are the sites where Essentras sells placements, not a random sample of the web.

Beyond the index

llms-full.txt and markdown pages

llms-full.txt is the file people confuse most with llms.txt. It is not in the proposal.

llms-full.txt is a convention from documentation platforms. Mintlify describes it as your whole documentation site in 1 file, each page with its title, source URL, description and full markdown (Mintlify docs, updated 10 Oct 2026). Mintlify also serves copies under /.well-known/.

Version 1 of the proposal described a tool that expanded llms.txt into a context file. Version 2 dropped that tooling: the proposal now expects agents to read llms.txt and follow its links.

Full files get big. On 4 Nov 2026, Anthropic’s developer docs served an llms.txt of 93,965 bytes and an llms-full.txt of 49,387,216 bytes, about 526 times larger. Mintlify splits generated llms.txt indexes above 100,000 characters into smaller files, and that limit does not apply to llms-full.txt.

Our view: a business site rarely needs llms-full.txt. Documentation and API sites can offer it next to llms.txt. Markdown copies of key pages do the same job page by page, and version 2 of the proposal shows agents how to find them.

File sizes in the wild, read on 4 Nov 2026

File	Size
llmstxt.org/llms.txt (the proposal’s own)	637 bytes
Median of 201 media sites in our survey	3,243 bytes
developers.openai.com/llms.txt	6,201 bytes
essentras.com/llms.txt (ours)	82,037 bytes
platform.claude.com/llms.txt (Anthropic)	93,965 bytes
platform.claude.com/llms-full.txt (Anthropic)	49,387,216 bytes (49.4 MB)

Let agents find your llms.txt and markdown pages (version 2)

When to use it: Add the HTTP header in your server or CDN settings, or the 2 link tags to each page. Replace the paths with your own.

HTTP Link header

Link: </docs/page.html.md>; rel="alternate"; type="text/markdown", </docs/llms.txt>; rel="describedby"

HTML link tags

<link rel="alternate" type="text/markdown" href="/docs/page.html.md">
<link rel="describedby" href="/docs/llms.txt">

From the llmstxt.org proposal, version 2. Optional: the file works without them.

Publishing

Where to host llms.txt, and how to test it

The file is plain text at a fixed address. Most problems in our survey came from the server, not the file.

How to publish an llms.txt file

01Save it as plain textName the file llms.txt and save it in UTF-8.
02Put it at your site rootIt should load at https://yourdomain.com/llms.txt. A file at a path such as /docs/llms.txt covers only the pages under /docs/.
03Serve it as text with status 200Use text/plain or text/markdown. In our survey, 323 of 327 files were text/plain.
04Keep it publicNo login, no firewall challenge and no redirect to another page. Lighthouse flags a server error on this file.
05Test itOpen it in a private window, then run the 2 commands below: status 200, a text content type, and your H1 on line 1.
06Review it on a scheduleUpdate it when prices, products or important URLs change, and note the review date in the details.

By platform

Platform	How
Any static host or framework	Put llms.txt in the folder served at the root, such as public/ in Next.js. essentras.com serves its file that way.
WordPress	Yoast SEO and AIOSEO can generate one, and so can Rank Math. Check the top of the output: in our survey, 73 plugin-made files had a “Generated by” line above the H1.
Wix	Generates an llms.txt for every Wix site, per the integrations list on llmstxt.org.
Mintlify and GitBook (documentation)	Generate llms.txt automatically. Mintlify also serves llms-full.txt and copies under /.well-known/.
A firewall or bot protection	Check that your bot rules let through the agents you want to read the file.

The platform list comes from the integrations section of llmstxt.org, read 4 Nov 2026; Rank Math from the files in our survey.

Test your llms.txt from the command line

When to use it: Replace example.com with your domain.

Status and content type (expect 200 and text/plain or text/markdown)

curl -sI https://example.com/llms.txt

First lines (expect "# Your site name" on line 1)

curl -s https://example.com/llms.txt | head -n 5

Avoid these

Common llms.txt mistakes

Each row comes from our survey, a source above, or the reference parser’s code.

Mistake	What goes wrong	Fix
Expecting rankings or AI citations	Google Search ignores the file, and no AI company documents using it in answers. In Ahrefs’ May 2026 data, 97% of files got no request at all.	Publish it as a convenience for agents. Work on the pages AI answers cite.
Writing robots.txt rules in it	27 files in our survey held User-agent and Disallow lines. llms.txt controls nothing.	Put crawler rules in robots.txt; keep llms.txt for the map.
A line above the H1	73 plugin-made files start with a “Generated by” line, while the format starts with the H1.	Remove the line, or change the plugin’s template.
An HTML page at /llms.txt	214 sites answered with an HTML page and status 200, so an agent gets markup instead of a map.	Serve a real text file, or let the address return 404.
Prose under H2 headings	The reference Python parser stops with an error at the first line under an H2 that holds no "- [title](url)" link.	Put prose above the first H2, or in the linked pages.
Sub-headings inside sections	Strict parsers read a ### line as the start of a new section.	Use separate H2 sections.
Relative or broken links	An agent follows the URLs as written, and the proposal’s examples use full URLs.	Use full https URLs and open each one.
A huge file	The proposal says the file should fit in an agent’s context. 18 files in our survey were over 100 KB; the largest was 21.7 MB.	Keep detail behind the links; use markdown pages or llms-full.txt for depth.
A file nobody updates	An agent that reads an old file gets old prices and dead links.	Review it when prices, products or URLs change, and date it.
Private or staging URLs	The file is public, so every link in it is public too.	List only pages you want anyone to read.

A live example

Essentras’s own llms.txt

Our file, read on 4 Nov 2026 at 18:17 UTC, and what our checker says about it.

Measure	essentras.com/llms.txt
Size	82,037 bytes, 282 lines, about 10,900 words
H1	Essentras
Summary blockquote	Yes: what Essentras is, how the brand is spelled, and that it is not Essentra plc
H2 sections	16, from “Who it is for” to “About the name”
Link items	137
Served as	text/plain; charset=UTF-8, status 200
Checker result	Valid, with warnings: 105 lines under H2 headings are not link items, and 8 sections hold no links
Reference Python parser	Stops at line 10, the first line under an H2 that is not a link item

Why it looks like this: our file doubles as a fact sheet. It answers what agents ask about us (prices, refunds, turnaround), points agents that buy on a user’s behalf to our API and MCP server, and spells out the brand name.

The trade-off: it is about 25 times the median file in our survey, and it puts lists and facts under H2 headings, which the format keeps for links. If you copy the approach, keep the prose above the first H2 or in linked pages, and strict parsers will read the whole file.

Open essentras.com/llms.txt

If you want help

Often added

Optional. Everything on this page works without us.

If you want help

Not ready to pay?

Start with the free checks.

FAQ

llms.txt: FAQ

What is llms.txt?▾

A proposed markdown file at the root of a website that gives AI agents a short map of it: an H1 with the name, a 1-line summary, then H2 sections of links with notes. Jeremy Howard proposed it on 3 Sep 2024, and version 2 is dated 10 Aug 2026. It is a proposal, not a web standard.

Does llms.txt help SEO or Google rankings?▾

No. Google’s AI optimization guide says Google Search ignores llms.txt files, so keeping one neither helps nor harms rankings, in AI Overviews and AI Mode too. Google added that note on 15 Jun 2026.

Does ChatGPT use llms.txt?▾

OpenAI does not say so: its crawler page covers robots.txt tokens only. In Ahrefs’ May 2026 log data, GPTBot was the AI bot that fetched llms.txt files most often, but a crawler fetching a file does not show that answers use it.

Do Claude and Perplexity read llms.txt?▾

Neither company documents it. Their crawler pages cover robots.txt only, though both publish llms.txt files for their own developer docs. Ahrefs saw Claude Code, Anthropic’s coding tool, among the top fetchers of llms.txt files, which fits the proposal’s main use: coding agents reading documentation.

Is llms.txt the same as robots.txt?▾

No. robots.txt tells crawlers what they may fetch, and AI companies document their robots.txt tokens, such as GPTBot and ClaudeBot. llms.txt grants and blocks nothing; it only describes your site. Put access rules in robots.txt.

Where do I put my llms.txt file?▾

At the root of your site, so it loads at https://yourdomain.com/llms.txt, as plain UTF-8 text with status 200. Version 2 of the proposal also allows a file under a path, such as /docs/llms.txt, which covers only the pages below it.

What is llms-full.txt?▾

A convention from documentation platforms, not part of the proposal: the full text of every docs page in 1 file. It can be huge (Anthropic’s was about 49 MB on 4 Nov 2026), so it suits documentation and APIs more than business sites.

Do I need an llms.txt file?▾

Not for Google. It takes a few minutes and helps agents that read it, mostly coding tools on documentation and API sites. Our view: if you sell services, spend your effort first on the pages AI answers cite, and add llms.txt as a tidy extra.

Is this llms.txt generator free and private?▾

Yes. It is free, needs no account and runs in your browser: nothing you type is sent to us or to any server, and the tool never visits the URLs you enter. Copy the result or download it as llms.txt.

How often should I update llms.txt?▾

Whenever prices, products or important URLs change. An agent that reads an old file gets old facts, so put the date of your last review in the details line.

Next steps

01Build the fileFill in the form or paste your URLs, then download llms.txt.Open the generator
02Publish and test itUpload it to your site root, then check the status, the type and line 1.Hosting steps
03Set your crawler rulesDecide in robots.txt which AI crawlers may read your site.AI crawler tokens
04See what AI says about you3 buyer questions asked live, free.Free AI check
05Track it every week7 assistants, free or $49 a month.AI Visibility Tracker

Related guides

Sources

Sources, read on 4 Nov 2026

Every source checked on 4 Nov 2026

Company documentation changes often. Each page below was read on the date shown; dates in brackets are the page’s own.

llmstxt.org: The /llms.txt file, version 2 (Jeremy Howard; 3 Sep 2024, modified 10 Aug 2026) · Read 4 Nov 2026
llmstxt.org: Changes, version 2 (August 2026) · Read 4 Nov 2026
AnswerDotAI/llms-txt: llms_txt/core.py, the reference parser · Read 4 Nov 2026
Google Search Central: Optimizing your website for generative AI features (updated 10 Jul 2026) · Read 4 Nov 2026
Google Search Central: Latest documentation updates, 15 Jun 2026, “Clarifying guidance on llms.txt files” · Read 4 Nov 2026
Google: Google’s common crawlers, Google-Extended (updated 30 Oct 2026) · Read 4 Nov 2026
Chrome for Developers: Lighthouse llms.txt audit (updated 5 May 2026) · Read 4 Nov 2026
Chrome for Developers: Lighthouse Agentic Browsing scoring (updated 5 May 2026) · Read 4 Nov 2026
OpenAI: Overview of OpenAI crawlers · Read 4 Nov 2026
OpenAI Developers: llms.txt · Read 4 Nov 2026
Anthropic: Does Anthropic crawl data from the web, and how can site owners block the crawler? (updated 7 Apr 2026) · Read 4 Nov 2026
Anthropic developer docs: llms.txt and llms-full.txt · Read 4 Nov 2026
Perplexity: Perplexity crawlers (updated 29 Oct 2026) · Read 4 Nov 2026
Perplexity docs: llms.txt · Read 4 Nov 2026
Mintlify docs: llms.txt and llms-full.txt (updated 10 Oct 2026) · Read 4 Nov 2026
Ahrefs: We Analyzed 137K Sites: 97% of llms.txt Files Never Get Read (15 Jun 2026, updated 27 Aug 2026) · Read 4 Nov 2026
Essentras survey: /llms.txt on 3,840 media sites in our catalog · Read 4 Nov 2026
Essentras: essentras.com/llms.txt · Read 4 Nov 2026

What we could not verify

Whether any AI assistant uses llms.txt when it writes an answer. No company documents it, and server logs show fetches, not use.
Microsoft (Bing and Copilot), Meta and xAI: we found no primary page of theirs on llms.txt, so none is cited.
We did not run Lighthouse’s Agentic Browsing audit ourselves; its rules here come from Chrome’s documentation.
The files of the 929 catalog sites that refused our survey bot or failed.

Found an error or a newer rule? Tell us

Essentras sells AI visibility tracking, AI SEO work and media placements. No assistant’s answer can be bought or promised, by us or anyone. Company documentation changes often: check the pages linked above before you act.