How to create an llms.txt file without uploading your site map
llms.txt is a small Markdown file that sits at the root of a website and tells a language model what the site is and which pages matter. It was proposed in September 2024 by Jeremy Howard of Answer.AI, and revised as v2 in August 2026. Big documentation sites publish one — the proposal’s own examples include developer docs — and several platforms now generate one automatically. But it is a proposal, not a standard: no standards body maintains it, no AI company is obliged to read it, and support varies from tool to tool.
This tool builds the file where your data already is: on your own device. When you paste a page list or a sitemap, your browser:
- Reads your pages — one per line as “Title | URL | description”, a bare URL, or the raw XML of your sitemap.xml. Titles are derived from the URL path when you do not supply one, and relative paths are resolved against the site URL you type in.
- Groups them into sections — by the first part of each URL’s path, so /docs/… pages become a “Docs” section and /blog/… pages become “Blog”. Turn grouping off for one flat list.
- Writes both files — llms.txt, and llms-full.txt with a short excerpt after each link — then checks the result: a missing H1, a non-absolute URL, a duplicate link, a link with no description, or a file that has grown past 100 links.
What makes this generator different
The established generators are either server-side crawlers — they fetch your site, read up to a few dozen pages, and rate-limit you — or command-line tools you install and run. The free browser generators that do exist take a form and write the file, but they cap the pages, hide the format rules, or leave you to guess what is wrong when the output is malformed. None of them, as far as we could find, combine all four things this page does: sitemap XML pasted straight in and parsed on-device, a live editable preview of both files, a checks table that names every format problem in plain English, and no server, no sign-up and no rate limit at all. This page is that combination — free, fully client-side, with a live network-request counter so you can verify the privacy claim yourself.
When is an llms.txt worth the ten minutes — and when is it not?
It is worth it for a documentation site, a blog with a clear best-of list, or any site whose most useful pages are not obvious from the homepage: one text file, no code, no change to your existing pages. It is not worth it as an SEO move — Google says it ignores the file for Search, and there is no public evidence it earns citations. It is also not a fix for a site AI cannot read: if your pages are blocked to AI crawlers or load as an empty shell, a map of them helps nobody. Fix the access first, then add the map.
Honest limitations
A proposal, not a standard
llms.txt was proposed in 2024 and revised in 2026. No standards body maintains it and no AI company has said its crawlers read third-party files. Publishing one is voluntary and cheap — not a guarantee of anything.
No ranking or citation claim
Google says it ignores llms.txt for Search, including AI Overviews and AI Mode. Nothing here promises traffic, rankings or mentions — only a correctly formatted file.
It never crawls your site
The tool works only on the text you paste. It cannot check that your URLs exist, fetch your page titles, or write excerpts for llms-full.txt. Check your links before publishing.
Descriptions are yours to write
A link without a description is valid but barely useful — and a sitemap gives no descriptions at all. The best files are curated by hand: pick your most important pages and say, in one line each, why they matter.
Small by design
The proposal says the file should fit in context. A 500-page dump of your sitemap is not a map, it is a sitemap in another format — and this tool warns you past 100 links for exactly that reason.
It grants and blocks nothing
llms.txt is a guide, not a gate. It does not replace robots.txt, it does not block crawlers, and it does not give any model permission it did not already have.