LLMs.txt Generator for Your Website
Turn your home page links or your XML sitemap into a clean llms.txt draft that points AI assistants to the pages that explain your site best.
What goes into the generated file
The file follows the llmstxt.org layout: a heading with the site name, a short quoted summary, then Markdown link lists under section headings.
- Site name
- Taken from the home page's og:site_name, then its Open Graph title, then its title tag, and the bare hostname if none of those exist.
- Site summary
- The home page meta description becomes the quoted line under the name. With no description, the line is left out rather than invented.
- Source and language line
- The site address plus the lang attribute of the home page, so a reader knows which version of the site the list describes.
- Page list from links
- In home page mode, up to 60 links on the same domain, with query strings, fragments and file links such as PDFs, images and scripts removed.
- Page list from the sitemap
- In sitemap mode, sitemaps named in robots.txt are tried first, then /sitemap.xml and /sitemap_index.xml. A sitemap index is followed into up to 8 child sitemaps, for up to 500 page addresses.
- Real titles and descriptions
- The first 25 pages after the home page are opened, three at a time, to read their actual title and meta description. Pages beyond that get a readable title made from their URL slug.
- Sections by path
- Each page is filed under the first folder of its URL, so /blog/post lands in Blog. Home comes first, common sections such as Docs, Guides, Blog, Products, Pricing and About follow in a sensible order, and the rest go alphabetically.
- Safe Markdown
- Square brackets are taken out of titles and parentheses and spaces in addresses are encoded, so an odd page title cannot break a link or the file.
How to create an llms.txt file
Enter your website
Any address on the site works. The generator always starts from the root of that domain.
Choose where pages come from
Pick the home page's links for a small, curated list, or the XML sitemap when you want every published page considered.
Run it and check the stats
You get the number of entries, sections, the file size and how many real titles were read.
Edit the draft
Remove thin or private pages, rewrite vague titles, and add one line of context to the links that matter most.
Publish it at the root
Download the file and upload it so it opens at yourdomain.com/llms.txt, the address shown under Publish At.
The home page gets about 12 seconds and each other page about 8 seconds, with up to 2 MB read per page. If the home page fails, or sitemap mode finds no sitemap with pages, the run stops with the reason.
Understanding the output
- Entries
- How many page links the file holds. A count much lower than your real page count usually means the home page links to few pages, so try sitemap mode.
- Sections
- How many path groups were found. One huge section with everything in it often means the site uses flat URLs with no folders.
- Titles Read
- Pages whose real title was fetched. Entries past the first 25 show a title built from the slug, which you may want to polish by hand.
- File Size
- The size of the finished text. Keep it lean: a short, well chosen list is easier for a model to use than a dump of every tag page.
- Publish At
- The exact address where the file should live. A copy in a subfolder will not be found.
- Built From
- Which source was used, and in sitemap mode the sitemap address that supplied the pages.
Common llms.txt mistakes and how to fix them
Sitemap mode finds nothing
Fix: Add a Sitemap line to robots.txt pointing to your real sitemap, or place it at /sitemap.xml. Sitemaps on another domain are ignored on purpose.
Titles that read like slugs
Fix: Pages outside the first 25 get a title made from the URL. Replace those with the page's real name before you publish.
Login, cart or thank you pages in the list
Fix: Delete them from the draft. The file should describe content worth reading, not app screens or checkout steps.
Everything lands in one section
Fix: Regroup the links by topic by hand, with clear section names such as Guides, Pricing or Support.
The summary line is missing
Fix: Your home page has no meta description. Write one plain sentence about what the site offers and add it to both the home page and the file.
Who should publish an llms.txt
- Software and SaaS teams with documentation they want AI assistants to quote correctly.
- Publishers who want their key guides, not their tag archives, to be the first thing a model sees.
- Agencies adding an AI readiness item to a technical SEO checklist.
- Small business owners who want a plain map of their services and pricing pages.
What llms.txt can and cannot do
llms.txt is a proposed standard, not a ranking signal and not an access rule. Each AI service decides for itself whether to read it. It does not block or allow any crawler: that is still the job of robots.txt.
LLMs.txt Generator FAQ
Short answers about the LLMs.txt Generator.
A Markdown file at the root of a site that lists its most useful pages with short descriptions, following the format proposed at llmstxt.org, so language models can find the right content quickly.
Home page links give a short list of what you already feature, up to 60 pages. The sitemap covers up to 500 pages and suits larger sites, but needs more trimming afterwards.
Only the first 25 pages are opened to read their real title. The others get a readable name built from the last part of their address.
At the root of your domain, so it opens at https://yourdomain.com/llms.txt. On most hosts that is the same folder as robots.txt.
No. robots.txt tells crawlers what they may fetch. llms.txt only suggests what is worth reading. Use both.
No, it is a draft and says so inside the file. Review the links and wording before you publish it.
No. Only addresses on the exact origin you entered are included, so links to a blog on a separate subdomain need a separate run.
Related SEO tools
AI Crawler Access Check
Which AI and search crawlers can reach your pages.
1 free run a dayXML Sitemap Extractor
Every URL inside a sitemap or sitemap index.
1 free run a dayRobots.txt Analyzer
28 checks on a live robots.txt file.
1 free run a daySchema Generator
Valid JSON-LD for 17 schema types from a form.
Free, unlimited