llms.txt Generator
Builds the llms.txt file that helps ChatGPT, Claude and Perplexity understand your site correctly: a short description, your key pages and contact details.
- Sign-in requiredMembership is free
- Open on the Pro plan100 runs a day
- Saved to your historyReachable from your account
What does this tool look at?
Anything that is easy to get wrong by hand, generated for you: meta tags, schema, robots.txt, sitemaps and redirect rules.
- Markdown output following the llms.txt standard
- A section for your key pages
- Instructions for uploading it to the root
What this tool measures
When a language model lands on your site it has to strip menus, cookie banners, sidebars and JavaScript-rendered blocks before guessing what you do. llms.txt removes the guesswork: you write who you are, what you sell and which pages matter, in plain Markdown.
The tool assembles the file in the order the format expects: an H1 title, a one-sentence summary as a blockquote line, a list of key pages, contact details and a notes section. Your site address and sitemap URL are added to the notes automatically.
Page lines are written as “Title | address | short description”. A relative address such as /services is joined to your domain and written out in full. Up to 100 pages are listed.
How to use it
- 1
Fit the description into one sentence
The check expects at least 40 characters. “Software company” says nothing; “Team building appointment, bookkeeping and CRM software for small businesses in İzmir” is enough for a model to describe you correctly.
- 2
List at least three pages
Fewer than three raises a warning and none is an outright failure. Start with service pages, pricing and contact, then add the three or four most-read articles rather than the blog archive.
- 3
Add a short note to each line
The note is optional but does the real work. “Services: web design, custom software, SEO and ad management” helps a model pick the right page far more than the bare word “Services”.
- 4
Include contact details
An email address or the URL of your contact page. Leaving it empty raises a warning, because a model recommending you then has nowhere to send the reader.
- 5
Publish it at the root
Placement details are below. After uploading, run the llms.txt Checker to confirm the file is reachable and its structure parses.
When to use it
- When your product names are common words and models keep confusing you with something else.
- After launching a new product or service, to announce it instead of waiting for AI tools to infer it from HTML.
- When most of your site sits behind a login and only a few pages are readable from outside.
- On documentation sites, to make clear which version is current.
Reading the result
Where the file goes
At the domain root, reachable at example.com/llms.txt. The server must send it as text/plain; some servers mark .txt as a download, and then nothing reads it.
In a Next.js project
For fixed content, place it at public/llms.txt. When the page list comes from a database, write a GET handler in app/llms.txt/route.ts that returns the text with a Content-Type: text/plain header, so the list refreshes on every deploy.
In WordPress
Upload the file to the root folder, the one holding wp-config.php, over FTP or the file manager. Do not create it as a post or page: content living at a URL like /llms-txt does not count as the standard location.
It must agree with robots.txt
If you block AI crawlers in robots.txt, nothing will ever fetch this file. Decide both together: publishing llms.txt belongs to the same policy as allowing AI bots.
Keep the file alive
Update it whenever a listed page is retired or moves. An llms.txt full of dead links leaves a worse impression than no file at all.
Questions about the llms.txt Generator
No. It is a community format proposed in 2024 with no role in Google's indexing pipeline. Even so, more AI tools look for it every month, and publishing one costs you nothing.
Not necessarily. llms.txt is a map; llms-full.txt is the plain text of your entire site. It makes sense for documentation sites and is overkill for a ten-page corporate site.
The file is a summary, not a permission slip. Training access is governed by robots.txt. If you do not want your content used for training, block those bots there; llms.txt will still help the search-oriented crawlers.
Five to twenty suits most sites. The goal is not to dump the site but to point at the pages a model should open when describing you. A long list dilutes the important entries.
Use it alongside these
The other tools in the same category. Run them together and you see the whole picture.
The tool found the problem. Now for the fix
We run technical SEO, content and AI visibility work end to end. Let us review your site and come back with a clear list of what to do.
Monday – Saturday 09:00 – 18:00