Free llms.txt Generator
Enter your website and get an llms.txt you can edit: your main pages, grouped into sections, each with a one-line description. Free, with no account. What is llms.txt?
Generated, not installed yet. This page can’t see your website, so step 3 checks the installed file.
Writing these by hand? AI-written page descriptions, from each page’s own text, are coming in paid plans. See the planned plans
Optional: the full-text companion
llms-full.txt puts the text of your pages in one file, for tools that read whole pages.
llms.txt Example output# Riverbend Ceramics > Hand-thrown stoneware from a small Bristol studio. ## Shop - [Mugs](https://example.com/mugs): Six glazes, 350ml. - [Bowls](https://example.com/bowls): Thrown to order.
A made-up shop. Your file lists your own pages; see complete examples.
- Free, no account
- Up to 50 pages a run, 3 runs a day
- Website, sitemap or URL list
- Follows your robots.txt
- Keeps no copy of your pages
What does the output look like?
An llms.txt file is plain Markdown: a title, a short summary, then sections of links, each with a one-line description. The example beside the generator has every part:
| Part | Looks like | What it is for |
|---|---|---|
| Title | # Riverbend Ceramics | The name of the site or organisation. One per file. |
| Summary | > A small studio in Bristol… | What the site is and offers, in a sentence or two. |
| Sections | ## Shop | Groups of related pages. The generator groups them by URL path; rename them if a clearer word fits. |
| Links | - [Stoneware mugs](https://…): Six glazes… | One page each, with a description after the colon. The generator drafts it from the page’s own title and description; rewrite the important ones with facts. |
| Optional | ## Optional | Background pages such as history or legal. A convention, not a rule: since llms.txt V2 it has no special behaviour. |
The description after each link does most of the work. It is what a model reads when deciding whether a page answers the question in front of it.
How does the llms.txt generator work?
Generate
Give it a website, a sitemap or a list of URLs. It reads your pages one at a time, following your robots.txt, and leaves out archives, feeds, search, media and account pages.
Review and download
Fix anything that is wrong, above all the descriptions of your most important pages, then download llms.txt. Pages the generator could not describe are flagged, so you know which lines to write.
Install it
Upload the file so it loads at https://yourdomain.com/llms.txt as plain text.
Where the file goes on your platform
| Platform | Where the file goes |
|---|---|
| Most hosting with cPanel or FTP | The web root, usually public_html/ |
| WordPress | The folder with wp-config.php, or generate it (see the WordPress guide) |
| Next.js, React (Vite), Astro, Nuxt | The public/ folder, or a route (see the Next.js guide) |
| Laravel, Symfony | The public/ folder |
| Netlify, Vercel, GitHub Pages | The folder your site is published from, such as public/, dist/ or the repository root |
| Shopify | Not possible directly; use a redirect or a proxy (see the Shopify guide) |
| Squarespace 7.1 | No upload: paste the text into the built-in setting (see the Squarespace guide) |
| Wix | Wix makes one for you; edit or replace it in the dashboard (see the Wix guide) |
The root is where every reader looks. Since llms.txt V2 a file in a subdirectory, such as /docs/llms.txt, is also valid and covers the pages under it, but keep one at the root too.
Check the installed file
Run its address through the validator, which also checks the content type and redirects. It reads the file as it is on your site now; nothing here can see your server, so this check is the only way to know it is installed.
Publishing the file doesn’t guarantee that any AI tool reads it. Whether one does is up to each AI company.
Coming soon: AI-written page descriptions and summaries, in paid plans. Everything on this page stays free. See the planned plans
Guides for the next step
Put it on your platform
Check and extend it
Frequently asked questions
What does an llms.txt generator do?
It writes the file for you. It finds your sitemap, reads each page’s title and description, groups the pages into sections and formats everything as the Markdown the llms.txt proposal describes. Writing a file by hand takes twenty minutes or more; a first draft from the generator takes about a minute, and you then edit the descriptions that matter most.
Is it llms.txt or llm.txt?
llms.txt, with an s. That is the name the proposal gives the file and the address AI tools request, so a file saved as llm.txt or llms.text will not be found. The generator always names the download correctly; upload it unchanged to your domain root.
Is this really free?
Yes. All the tools are free and there is no account to create. Optional paid AI features are planned, and the free tools stay free when they arrive.
Do you store my site content?
No. Pages are fetched, turned into the output file and discarded when the request ends, and the file itself is not kept. The server keeps only operational records: short-lived pseudonymous counters for the free limits, daily usage totals and error logs. See the privacy page.
How many pages should I include?
For most sites, 30 to 100 well-chosen URLs beat an exhaustive dump. The file is a curated index, not a second sitemap: if everything is listed as important, nothing is. A free run reads up to 50 pages, so for a larger site start with the pages that matter most (paste them in the URL list tab) and add the rest by hand.
Will adding llms.txt improve my search rankings?
No. Google says you do not need AI text files to appear in its AI features, and there is no evidence of any ranking effect. The file is about how accurately models describe you when they do read your site. The longer answer is here.
How often should I regenerate it?
After any change that alters what your site is for: a new product area, a restructure, a large batch of pages. For a stable site, a check every month or two is plenty.
Do AI companies actually read this file?
Some do, unevenly. The proposal’s V2 notes say coding agents use these files reliably; general assistants are less consistent. Treat it as cheap insurance: a few minutes of work, and no downside if a given model ignores it. Whether it is worth it for you.