resulthack

llms.txt validator: check your file line by line

Paste the llms.txt you already have, hand-written or made by a CMS plugin, and see each line checked against the llmstxt.org proposal (v2, read 24 Sep 2026). It runs in this tab: nothing you paste is sent anywhere. No file yet? Build one in the generator.

The format, as the proposal gives it

The llms.txt proposal by Jeremy Howard (llmstxt.org, first published 3 Sep 2024, revised as v2 in August 2026, read 24 Sep 2026) is a Markdown file whose parts come "in the specific order":

  1. an optional byte-order mark (BOM);
  2. an H1 with the name of the site or project, "the only required section";
  3. a blockquote with a short summary;
  4. zero or more Markdown sections "of any type except headings" with more detail;
  5. zero or more H2 sections, each a "file list": a Markdown list whose items hold a required [name](url), then optionally a : and notes.
# Acme Tents

> Ultralight tents and tarps, sold by mail order across the EU.

Prices are in euros and include VAT.

## Products

- [Ultralight 2](https://example.com/tents/ultralight-2.md): two-person tent, 1.1 kg

## Optional

- [Company history](https://example.com/about/history.md)

An H2 named "Optional" holds secondary links an agent can skip. Since v2 that is a convention only; the proposal's changes page says it no longer carries "mechanical semantics". v2 also lets a file sit at a sub-path, such as /docs/llms.txt, covering the pages under that path, and it suggests rel="describedby" links pointing to it.

Errors: the file does not follow the format

Warnings: allowed, or not addressed, but probably not what you meant

Info

The number of characters, lines, links and sections; a BOM; CRLF (Windows) or bare CR line endings; trailing spaces; no summary; a missing final newline. None of these break the format. Line endings and a BOM survive only when you drop or open a file: a pasted text box always holds LF endings.

What the proposal does not say

The checker does not invent rules for these. Where it warns, the warning says why.

The cleaned version

It changes four things and nothing else: * and + bullets become -; runs of blank lines become one, with one blank line around each # heading; trailing spaces and tabs are removed; line endings become LF. Fenced code is left alone apart from trailing spaces, and a BOM is kept. It does not fix errors, reorder anything or touch your wording. Cleaning a cleaned file changes nothing. The test suite runs the generator's output for three sets of choices (a name only; a full file with summary, details, two sections and Optional; names full of * _ [ ] and backticks) through this checker: each comes back unchanged, with no errors and no warnings.

Who reads llms.txt

Publishing an llms.txt is cheap, but no search engine or AI operator we checked says it uses one to rank, crawl or cite a site. What the sources say, each checked on 24 Sep 2026:

So a valid llms.txt is useful to an agent or a tool that goes looking for one, for example a coding assistant pointed at your docs. This checker tells you whether yours can be read the way the proposal describes. It makes no promise about traffic, ranking or citations.

Who publishes one

Of the 1,000 most-visited sites on Tranco list PY35J (generated 8 Oct 2026), checked on 9 Oct 2026, 118 serve an llms.txt (a 200 text response that is not an HTML page), and 52 of them pass this validator with no errors. 19 of those also have no warnings. 27 of the 118 also serve an llms-full.txt.

118

serve an llms.txt 11.8% of all sites

52

pass with no errors 44.1% of files

27

serve an llms-full.txt 2.7% of all sites

Why the others fail

Each of the 66 failing files is counted once per kind of error it has, so the rows can add up to more than 66. The names are the ones the validator's error list above uses. In the files we opened by hand, the usual cause of the first row is an H2 section written as a list of plain facts (- Founded in 2005) or bold labels rather than [name](url) links: Markdown reads them fine, but the proposal reserves H2 sections for lists of links.

ErrorFiles
A list item in an H2 section that is not a well-formed link53
More than one H19
The H1 is not the first line with text8
No H15

The highest-ranked sites that serve one

  1. cloudflare.com rank 2, has errors
  2. github.com rank 29, no errors
  3. fastly.net rank 36, has errors
  4. wordpress.org rank 49, has errors
  5. digicert.com rank 51, has errors
  6. adobe.com rank 62, no errors
  7. opera.com rank 80, has errors
  8. workers.dev rank 82, has errors
  9. sentry.io rank 96, has errors
  10. b-cdn.net rank 101, has errors

Each links to the site's live file, which may have changed since 9 Oct 2026. Paste one into the checker above to see its lines.

What the number cannot tell

We asked once, from one server, at the root of each domain. A site may keep its file at a sub-path (/docs/llms.txt, which v2 of the proposal allows) or on another host such as docs. or developers., and we did not look there; a site may also block or vary for an unknown User-Agent. "Serves" counts any non-empty text answer, so a plain-text "Not found" page sent with status 200 counts as a file and then fails for having no H1. Passing means the file follows the format, not that its links work or that any AI agent reads it. Of the rest, 115 answered /llms.txt with an HTML page, 349 with 404 or a similar "not here", 85 with 401 or 403, 102 failed (timeout, TLS or server error), 222 are infrastructure names with no website at the bare domain, 2 sent a file that is not text, 7 sent an empty file. The ranked list is Tranco's and is not republished here; we keep only these counts and the ten names above.

What this checker cannot see

Made with AI tools. · Privacy