Anobee

Free tool · SEO & AI Search

llms.txt Generator and Validator

An llms.txt validator reads the file the way a parser does and shows, line by line, what it understands and what it skips. Check any site's llms.txt against the v2 proposal, or fill in a short form to generate a valid file you can copy or download.

What to validate

A domain checks /llms.txt. For a file on a subpath, enter its full address, such as example.com/docs/llms.txt. Nothing is stored.

What llms.txt is, and what it is not

llms.txt is a Markdown file that gives AI agents a short, curated map of a site: what it is, and which pages to read for what. Jeremy Howard proposed it in 2024, and version 2 of the proposal was published in August 2026. Documentation platforms generate it automatically, and Chrome's Lighthouse checks for it in its agentic browsing audits.

It is not a ranking signal. Google's AI optimization guide says Google Search does not need the file and that it neither helps nor hurts rankings. It does not control crawler access either; robots.txt does that. The llms.txt implementation guide covers what the evidence does and does not support.

What the validator checks

The v2 proposal fixes the order of a file: an optional byte-order mark, then an H1, an optional blockquote summary, optional details in any Markdown except headings, and then H2 sections of links written as - [name](url): notes. The validator reads the file in that order and reports three kinds of finding:

  • Errors break the format, so a parser reads the file wrongly or not at all: a missing or late H1, a second H1, an empty file, a code block that never closes, or a URL that returns an HTML page instead of the file.
  • Warnings are content a parser silently drops: list items that do not start with a [name](url) link, plain text inside an H2 section, H3 headings, a summary that is not straight after the H1, relative or non-web links, empty sections, and a file with no links at all.
  • Tips are improvements: a missing summary, links without a description, http links, the same URL listed twice, and a file over 50,000 characters. The size limit is Anobee's rule of thumb; the proposal only asks that the file stays small enough to fit in context.

"Parser" here means the link-list shape the proposal and its reference parsers match: a list item that starts with a Markdown link. A language model reading the file as plain text may still understand a line a parser skips, but you cannot rely on it.

Example: anobee.com, checked on 28 October 2026

The first run of this validator on anobee.com's own file found no errors and one warning. The file had a title, a summary and 19 described links in four sections, but it ended like this:

## Optional

- [Blog archive](https://anobee.com/blog/): Every published article, newest first.
- [Sitemap](https://anobee.com/sitemap.xml): Machine-readable list of every indexable URL.
- [RSS feed](https://anobee.com/rss.xml): Subscribe to new articles.

This file is informational and does not guarantee AI citations or rankings.

The validator reported: line 36 is plain text inside the "Optional" section, so parsers ignore it. A parser reads only link lists inside sections, and the disclaimer sat after the last list. The fix was to move the sentence above the first H2, into the details part of the file. That is how the file is served now, and the Try anobee.com button shows the result.

Lighthouse had already caught a bigger mistake in the same file before this tool existed. Its entries were bare URLs after a label, so as Markdown the file contained no links at all, and Lighthouse said so. The validator reports that case as a warning too.

How to use the generator

  1. Enter the site name, which becomes the H1, and a one-sentence summary.
  2. Add a few sections of your most useful pages, each with a short description. Fifteen to thirty well-chosen links beat a copy of your sitemap.
  3. Put secondary links in a section named Optional, which agents may skip when context is short.
  4. Copy or download the file and publish it at /llms.txt, served as plain text. Then run the validator on the live address to confirm the server returns the file and not an HTML page.

Already have a file? Validate it, then choose Fix it in the generator to load it into the form. Text a parser would skip is not carried over, which is exactly the content worth moving.

What the result cannot tell you

  • Whether anything reads it. Check your server logs for requests to /llms.txt. A valid file that nothing fetches does nothing.
  • Whether the links work. The validator checks each link's form, not whether the page loads. It does not fetch them.
  • Markdown versions of pages. v2 recommends .md versions of key pages and rel="alternate" and rel="describedby" links to them. Those are not checked.
  • Crawler access. A perfect llms.txt is useless to a crawler your robots.txt blocks. Run the AI Crawler Access Checker as well.

What to do next

Frequently asked questions

Does llms.txt help SEO or Google rankings?

No. Google's AI optimization guide says Google Search does not need llms.txt, and that the file will neither help nor hurt your visibility or rankings. OpenAI's and Anthropic's crawler documentation covers robots.txt, not llms.txt. The file helps agents and tools that choose to read it find your best pages, so treat it as cheap infrastructure, not a ranking tactic.

Who actually reads llms.txt?

Mostly coding agents and documentation tools. The proposal itself says the files are used most heavily for software documentation, where agents follow them to API references and tutorials. Your server logs are the only way to see whether anything reads yours: look for requests to /llms.txt and note the user agents.

Where do I put llms.txt?

At the root of the site, so it loads at example.com/llms.txt. Version 2 of the proposal also allows a file on a subpath, such as /docs/llms.txt, which covers only the pages under that path. When more than one file applies, agents should use the most specific one.

Do I need llms-full.txt as well?

No. The v2 proposal is built around one short llms.txt that links to clean Markdown versions of your pages. A single large file with all your content is not part of it. Publish one only if a tool you rely on asks for it.

Is llms.txt the same as robots.txt?

No. robots.txt says which crawlers may fetch which pages. llms.txt is a reading guide for agents that are already allowed in. It cannot block or allow anything, so check crawler access separately with the AI Crawler Access Checker.

Which parts of the file are required?

Only the H1 with the site or project name. The summary blockquote, the details and the H2 link sections are optional, but a file with no links gives an agent nowhere to go.

Last updated 28 October 2026 by Bibek Thapa, against the llms.txt v2 proposal. Privacy: pasted and generated files never leave your browser. A website check asks Anobee's server to fetch that one llms.txt address, with the user agent AnobeeLlmsTxtChecker/1.0; neither the address nor the file is stored.