llms.txt Generator

Runs in your browser · nothing is uploaded

Format: llmstxt.org

Site

The H1 and the summary agents read first

Site nameH1 · required

OriginScheme + host · required

Summary0 / 320

DetailsOptional · plain Markdown, no headings

Sections

Each becomes an H2 with a list of links

No sections yet

Add a section to start listing links

Most files have a Docs section, and an Optional section for links agents can skip.

Quick add

llms.txt

Waiting for input

# Your site name > A one- or two-sentence summary of what the site is and who it is for.

Make agents find it

Serve the file at the root of your origin. Then point to it from the places crawlers and agents already read.

1 · Publish at

https://example.com/llms.txt

2 · robots.txt

User-agent: GPTBot
Allow: /

User-agent: ClaudeBot
Allow: /

User-agent: PerplexityBot
Allow: /

User-agent: Google-Extended
Allow: /

User-agent: Applebot-Extended
Allow: /

Allows GPTBot, ClaudeBot, PerplexityBot, Google-Extended and Applebot-Extended by name. Only add it if you want those crawlers on your site.

robots.txt options: Content-Signal and Sitemap

Content-Signal (optional)

Keys: search, ai-input, ai-train; values: yes or no. A proposal, not a standard; crawlers may ignore it.

Sitemap URL (optional)

3 · HTTP header

Link: </llms.txt>; rel="describedby"; type="text/plain"

Send it with your HTML pages, set in your server or host configuration.

4 · <head> of every page

<link rel="describedby" href="/llms.txt" type="text/plain">

Add a site name, origin and summary to enable download.

Why Use the llms.txt Generator?

An llms.txt file is a short Markdown index at the root of a site: a heading, a one-paragraph summary, and lists of links to the pages that matter, each with one line saying what is on it. It is a proposal published at llmstxt.org, meant for language models and agents that would otherwise have to read your navigation, scripts and footers to find the useful pages. The format is small but strict: one H1, a blockquote right under it, H2 sections of "- [title](url): description" lines, and an Optional section last. This generator builds that file from a form, resolves root-relative paths against your origin, escapes characters that would break a Markdown link, and warns about the mistakes that are easy to ship: links on another host, the same URL listed twice, empty descriptions and titles copied from a template. It also gives you the robots.txt group, HTTP Link header and HTML link tag that point to the file.

llms.txt is a plain Markdown file served at /llms.txt that describes a website for language models. The format is defined at llmstxt.org: an H1 with the site or project name, a blockquote with a short summary, optional paragraphs of context, and then H2 sections, each holding a list of links written as - [title](url): description. A section named Optional is special: it holds links an agent may skip when its context window is short.

This generator writes the file from a form. The site name becomes the H1, the summary becomes the blockquote, and the Details text is placed under it as prose; a Markdown heading typed into Details is escaped, because a heading there would start a new section. Each section becomes an H2 in the order you set, empty sections are left out, two sections with the same heading are merged, and a section called Optional is moved to the end. Link URLs may be absolute (https://…) or root-relative (/guides/); relative ones are resolved against the origin. Square brackets in titles and parentheses in URLs are escaped so each line stays a valid Markdown link.

The page checks the file as you type. Errors block the download: an origin that is not an absolute http(s) URL, a link without a title, and a URL that is neither absolute nor root-relative or still holds an <angle-bracket> placeholder. Warnings do not block it: a link on another host, an http link on an https site, the same URL listed twice, a missing description, a title left over from a template, a summary longer than 320 characters and a file larger than about 100 KB. URLs are checked for shape only; nothing is fetched, so the page cannot tell whether a linked page exists.

Next to the file you get four things to publish: the address the file belongs at, a robots.txt group that allows the main AI crawlers by name (with an optional Content-Signal line and Sitemap line), an HTTP Link header and an HTML <link> tag that both point to /llms.txt. To write the rest of robots.txt, use the robots.txt Generator, and to check whether a page is ready for AI search, use the GEO Analyzer.

How it compares

robots.txt, sitemap.xml and llms.txt answer different questions. robots.txt tells crawlers which paths they may fetch, and the robots.txt Generator writes it. sitemap.xml lists every URL you want indexed, with dates, and the XML Sitemap Generator builds it. Search engines have long-standing, documented support for both. llms.txt is a short, human-written index of the pages worth reading, with a sentence about each, for language models. It is a proposal from llmstxt.org rather than a standard: no search engine or AI company is known to guarantee that it reads the file, and there is no evidence that it changes search rankings. It costs little to publish and does not replace the other two files.

Tips for a Useful llms.txt

List the pages that answer questions, not every URL. A sitemap already lists everything; llms.txt is the short version with a sentence per link.
Write each description as the reason to open the page, for example "Every queue option, with its default value", rather than repeating the title.
Keep the file small. The generator warns above about 100 KB; move detail into the linked pages.
Use absolute https URLs or /paths/ on your own origin. A link on another host is allowed but flagged, in case it was a mistake.
Put the changelog, legal pages and archives under Optional, so an agent with little context space can skip them.
After publishing, check that https://your-site/llms.txt returns status 200 with a text content type, and run the GEO Analyzer on your homepage.

Frequently Asked Questions

What is an llms.txt file?

It is a Markdown file at the root of a website, for example https://example.com/llms.txt, that gives language models a short index of the site: a name, a one-paragraph summary and lists of links to the important pages, each with a one-line description. The format is a proposal published at llmstxt.org.

Where do I put llms.txt?

At the root of the origin, next to robots.txt, so it is served at https://your-domain/llms.txt with status 200 and a text content type such as text/plain or text/markdown. The Publish at box shows the exact address for the origin you entered. A file in a subfolder is not where agents look for it.

What format does llms.txt use?

Plain Markdown in a fixed order: one H1 with the site name, a blockquote summary right under it, optional paragraphs without headings, then H2 sections that each hold a list of links written as - [title](url): description. There is no front matter and no HTML. This tool writes that order for you and escapes characters that would break a link.

What is the Optional section?

The llmstxt.org format gives the H2 heading Optional a special meaning: links under it are secondary and can be skipped by an agent that has little context space. Changelogs, legal pages and archives fit there. The generator always writes Optional last, even if you add it first, and merges two Optional sections into one.

How do AI agents find llms.txt?

Tools that support the proposal request /llms.txt at the root of the site directly. You can also point to it: the page gives you an HTTP Link header and an HTML link tag with rel="describedby" that reference /llms.txt, and a robots.txt group that allows the main AI crawlers by name. None of these guarantee that a given agent reads the file.

Does llms.txt improve my Google ranking?

There is no evidence that it does. Google has not said it uses llms.txt for ranking, and the file is not part of its documented crawling or indexing. Treat it as a convenience for language models and agents that choose to read it, not as a search engine optimisation lever.

Does this tool build llms-full.txt?

No. llms-full.txt is a related convention where the full text of the linked pages is concatenated into one file. This tool writes only the llms.txt index, which links to your pages instead of copying them. It does not fetch your pages, so it could not build the full version without uploading or crawling anything.

What are content signals in the robots.txt snippet?

Content-Signal is a proposed robots.txt line that states how you allow collected content to be used, with three keys: search, ai-input and ai-train, each set to yes or no, for example search=yes, ai-input=yes, ai-train=no. It is optional, the page checks its syntax, and it is a statement of preference: crawlers are not required to honour it.

Is anything I type uploaded?

No. Runs in your browser; nothing you type is sent to FindUtils. Analytics and ads may load on the page. The draft is saved in your browser's local storage so a reload keeps it; Reset clears it. Links are checked for shape only and are never fetched.

Can I generate llms.txt from code or an AI assistant?

Yes. The same generator is available over the REST API with POST https://api.findutils.com/api/tools/llms-txt-generate/execute and as the MCP tool llms_txt_generate. Both take site_name, origin, summary, details, sections, optional_links, content_signal and sitemap_url, and return the file, the warnings and the three snippets.

Run it from code

$ curl -X POST \
  api.findutils.com/api/tools/llms-txt-generate/execute \
  -d '{"site_name":"Example","origin":"https://example.com","summary":"Example builds small tools for developers.","sections":[{"heading":"Docs","links":[{"title":…'

$ claude mcp add findutils \
  --transport http https://mcp.findutils.com/
# then call findutils:llms_txt_generate

API reference Add to Claude