Free llms.txt generator
- No signup
- Free
- 5 files a day
llms.txt is a proposed standard. Some AI tools and developer platforms read it, but major AI search engines have not confirmed they use it. It is cheap to add and will not hurt, but do not expect it alone to change your AI visibility.
llms.txt is a plain text file at the root of a site. It lists important pages with short descriptions so a tool that reads the file can see what the site covers. This generator builds that file from your pages. It will not, on its own, make ChatGPT recommend you.
What llms.txt is
The proposed format starts with a site name as a heading, a short summary in a blockquote, then sections of links with one-line descriptions. Optional links sit under their own heading. The point is curation. You choose the pages that explain the business.
It is not a sitemap. A sitemap can list thousands of URLs. llms.txt should stay short and readable. It is not robots.txt. robots.txt talks about fetching. llms.txt talks about meaning. Keep all three jobs separate.
Format explained with an example
The first line is a markdown H1 with the site name. The blockquote is one paragraph about what the business does. Sections use H2 headings such as Services or Guides. Each bullet is a markdown link, a colon, and a factual line of about twenty words.
This generator writes those lines from the page title, meta description, and opening text. It should not invent awards or cities that are not on the page. Edit any line that sounds like marketing fluff before you upload the file.
- # Site name
- > One-paragraph summary
- ## Services
- - [Page title](https://example.com/service/): what the page covers
- ## Optional
- - [Secondary page](https://example.com/note/): shorter context
llms.txt vs robots.txt vs sitemap
robots.txt decides which paths a crawler is asked not to fetch. A sitemap lists URLs you are willing to have discovered. llms.txt is a curated reading list with descriptions. Blocking a path in robots.txt and listing it in llms.txt is a contradiction. Fix one of them.
If AI crawlers you care about are blocked, an llms.txt file will not help those bots read the pages. Start with the robots.txt tester, then generate llms.txt for the URLs you want described.
Does it help AI visibility?
Be honest: major AI search engines have not confirmed that they use llms.txt. Some developer tools and early platforms read it. Adding the file is cheap and should not hurt when the descriptions are accurate. Treating it as a ranking lever is a mistake.
What still moves AI answers is clearer: a consistent brand name, pages that answer the questions people ask, and third-party mentions that match the site. Use the AI visibility checker to see the gap. Use this file as a tidy map, not as the strategy.
How to add llms.txt on WordPress, Shopify, and Next.js
On a normal host or WordPress with file access, upload llms.txt to the site root so it answers at /llms.txt. Some hosts need you to allow that extension. Fetch the URL in a private window after upload.
Shopify does not give every store a simple root text file. You may need an app, a proxy, or another hosting path. Do not pretend the theme editor always exposes /llms.txt. Next.js can place the file in public/llms.txt or serve it from a route handler. Prefer one method.
After it is live, open the file yourself. Check that every link returns 200 and that the descriptions still match the pages. When you change a service page, update the line. A stale llms.txt is worse than a short one.
How I pick the pages
Homepage, main services or products, pricing, about, and contact usually come first. Guides that answer real questions earn a place when they are not thin. Tag archives, pagination, and legal pages stay out unless you have a reason.
The discover step ranks candidates and lets you untick anything that should not be advertised. That control matters. A file that lists every blog tag teaches nothing. Twenty strong URLs beat two hundred weak ones.
Keeping the file honest over time
Regenerate when you launch a new service line or retire a page. A link to a 404 teaches a reader nothing. Prefer deleting a line over leaving a broken URL.
Do not stuff every blog post into the optional section. Pick the guides that still match how you sell. Seasonal posts can wait until they are current again.
If two pages say the same thing, keep the stronger URL. llms.txt is a place to remove duplicates, not to list them. The same discipline you use for canonicals helps here.
Write descriptions in the same language as the page. A Spanish service page should not get an English one-liner unless the page itself is bilingual. The model prompt asks for facts from the content. Still read the output.
When you care about AI answers, spend more time on entity consistency and the AI visibility checker than on polishing llms.txt punctuation. This file is a neat map. The map is not the territory.
Five generations a day is enough for edits. If you are iterating copy, edit the textarea locally and upload when the wording is stable. Burning the daily limit on tiny synonym changes helps nobody.
Tell your developer where the file lives. A marketing drive folder is not /llms.txt. The live URL must return text/plain or a clear text body at the site root, or the exercise was only a draft.
If the discover step finds almost nothing, the site may hide links behind JavaScript or block the fetch. Fix crawl access first. An empty candidate list is a crawl problem, not a reason to invent twenty URLs by hand.
I treat llms.txt like a receptionist script. Short, factual, and updated when the offer changes. It is not a place for adjectives, and it is not a substitute for the pages themselves.
When the draft looks right, download it, upload it, then fetch /llms.txt yourself. Only then move on to the visibility checker.
If a page is private, leave it out. llms.txt is for what you are happy to summarise in public. Password pages and staging hosts do not belong in the list. Keep the public story short and clear.