llms.txt generator
Enter a URL and get a ready-to-paste, spec-valid llms.txt — pages discovered from your sitemap, named from each page’s own title and description. Nothing invented.
Last updated
What is llms.txt?
An llms.txt file is a markdown index served at /llms.txtthat gives AI agents a curated, link-rich map of your site — similar to how sitemap.xml works for search engines, but optimised for LLMs. This generator builds the first draft for you.
How the generator builds your file
- Reads your homepage for the site name and one-line summary (Open Graph and meta tags).
- Discovers pages from
/sitemap.xml, or falls back to crawling homepage links. - Fetches each page for its own title and description — up to 50, the site’s actual text, nothing generated.
- Groups the links into H2 sections by URL path and emits a spec-valid file in the
[name](url): descriptionformat.
Does the output pass the llms.txt checks?
Yes, for the structural checks. The file is built to satisfy the seven content rules the llms.txt validator enforces (L2–L8 of the llmstxt.org spec). Two checks are publish-time steps only you can do, and one is an optional companion file — run the live file through the validator after you publish to confirm.
| Check | Requirement | Handled by |
|---|---|---|
| L1 | Served at /llms.txt (HTTP 200)Upload the file to your site root. | You, at publish |
| L2 | Starts with an H1 (project name)From the homepage title. | Generated |
| L3 | Valid, parseable MarkdownAlways well-formed by construction. | Generated |
| L4 | Blockquote summary below the H1Replace the placeholder if no summary was found. | Generated |
| L5 | H2 sections with link listsGrouped by URL path — reorder to taste. | Generated |
| L6 | Links as [name](url): descriptionDescriptions come from each page's meta. | Generated |
| L7 | Linked URLs are reachableOnly pages that returned HTTP 200 are listed. | Generated |
| L8 | ## Optional section used correctlyOmitted — it's not required. | Generated |
| L9 | Content-Type: text/plainServe the file as plain text, not HTML. | You, at publish |
| L10 | llms-full.txt companionA separate full-content file the generator doesn't build. | Optional companion |
After you generate it
Review the summary and sections, publish the file at your site root, then validate the live llms.txt against the 10-check llmstxt.org suite and run a full agent-readability scan.
Frequently asked questions
- What is an llms.txt generator?
- It's a tool that builds an llms.txt file for your site automatically. You enter your URL, and it discovers your pages (from your sitemap.xml, or by following homepage links if there's no sitemap), reads each page's title and meta description, and assembles a spec-valid llms.txt — an H1 site name, a blockquote summary, and H2 sections of [page name](url): description links — that you can copy, review, and publish at /llms.txt.
- Does it make up the page descriptions?
- No. Every link name and description comes from the page's own <title> and meta description (or Open Graph tags). Nothing is invented by an AI. If your homepage has no meta description, the summary blockquote is left as a clearly-marked placeholder for you to fill in, rather than fabricated.
- What if my site has no sitemap.xml?
- The generator falls back to crawling the links on your homepage to find pages. That's shallower than a sitemap, so you may want to add a few links by hand afterwards. If no pages can be discovered at all, you still get a valid starter template with your site name and homepage that you can extend.
- How many pages does it include?
- Up to 50, to keep generation fast and the file readable. llms.txt is meant to be a curated index of your most important pages, not a dump of every URL — so for large sites, treat the output as a strong starting point and trim or reorder it to highlight what matters most to an AI agent.
- What do I do after generating it?
- Review the file (especially the summary and section grouping), save it as /llms.txt at your site root served with a text/plain content type, then validate the live file with the llms.txt validator and run a full agent-readability scan to confirm it passes and to find anything else worth fixing.