llms.txt Generator

An llms.txt file is a short, plain-text summary of your website written for AI tools rather than for people or search engines. It sits at the root of your domain, at /llms.txt, and uses simple Markdown: a title, a one-sentence summary, a few notes, and then sections of links with a short description for each one. The idea is that when an AI assistant or an AI crawler wants to understand your site, it can read one small file instead of wading through navigation menus, cookie banners, scripts, and ads on hundreds of pages.

The format was proposed in late 2024 by Jeremy Howard of Answer.AI and is documented at llmstxt.org. It borrows its spirit from robots.txt and sitemap.xml: one predictable file in a predictable place. The difference is that robots.txt says what crawlers may not do, and a sitemap just lists URLs, while llms.txt explains what your site is and points to the pages that best represent it.

Less than you might think. This isn't a second sitemap, so you don't want every URL you own in it. A good llms.txt gives an AI the shortest path to understanding what you do and finding your best material. In rough order of importance:

• Your name and a one-line summary. Write the summary the way you'd describe the site to someone who has never heard of you: what it is, who it's for, and what makes it different.
• Any context an AI would otherwise get wrong. If your prices are starting points, your articles are opinion pieces, or your docs cover version 3 but not version 2, say so in the details.
• Your cornerstone content. The guides, docs, product pages, or service pages you'd be happiest to have quoted. Ten carefully chosen links beat two hundred random ones.
• A one-sentence description for each link. Say what's on the page, not just its title. "How to price a blog management retainer, with real numbers" is far more useful than "Pricing guide".
• Secondary pages under Optional. Legal pages, changelogs, and older announcements belong in the Optional section, which tells an AI it can skip them when it's short on room.

Skip anything you wouldn't want summarised out of context, like thin tag pages or half-finished drafts. And keep descriptions honest. An AI that reads your file and then finds the page doesn't match will trust the rest of it less.

In the top-level folder of your website, the same place as robots.txt, so that it loads at https://yourdomain.com/llms.txt. With most hosts that means uploading it over FTP or the file manager into public_html or your site's root directory. On WordPress, the easiest route is often a plugin: several SEO plugins, including Yoast and Rank Math, now have an llms.txt option, or you can upload the file next to your wp-config.php. On Shopify, Squarespace, and Wix you can't drop files into the root directly, so you'll need an app, a redirect, or a hosted copy on your own subdomain.

Once it's up, open the URL in a browser. You should see plain text, not a 404 page and not your site's theme wrapped around it. If you see a styled page, the server is treating the request as a normal page instead of serving the file.

They solve different problems and you can use all of them together.

• robots.txt is about permission. It tells crawlers, including AI crawlers like GPTBot and ClaudeBot, which parts of your site they may fetch. It doesn't describe anything. Our Robots.txt Tester and Generator handles that file.
• ai.txt and robots.json are newer, less widely adopted files for stating your terms for AI training and reuse. Our AI.txt / Robots.json Checker and Generator covers those.
• llms.txt is about understanding. It's a curated map of your site with context, written so an AI gets the right picture quickly. It grants no permissions and blocks nothing.
• llms-full.txt is an optional companion file that holds the full text of your key pages in one big Markdown document, so a model can read everything in one request. It's mostly used by documentation sites. This tool makes the standard llms.txt, which is the one to start with.

Honest answer: not on its own, and nobody can promise it will. llms.txt is a proposed standard, not a rule the big AI companies have agreed to follow. Google has said it doesn't use the file for AI Overviews, and OpenAI and Anthropic haven't documented support for it in their crawlers either. What does read it today is mostly developer tooling, coding assistants, and some AI agents and search startups, which is why it has caught on fastest with software documentation sites.

So why bother? Because the cost is close to zero and the downside is nil. The file takes twenty minutes to write, it doesn't affect your rankings or your page speed, and if adoption grows you're already in place. It's also a genuinely useful exercise: deciding which ten pages define your business, and writing a straight one-line description of each, tends to show you where your site is unclear. Think of it the way you'd think of schema markup in 2012. Not required, not guaranteed to pay off, but cheap insurance that occasionally turns into a real advantage.

What does get you cited in AI answers is the same thing that has always worked: clear, specific, well-structured pages that answer real questions and are linked to by other sites. llms.txt is a pointer to that work, not a substitute for it.

Yes. Open your sitemap in a browser (usually /sitemap.xml, /sitemap_index.xml, or /post-sitemap.xml), select all, copy, and paste it into the "Add Pages in Bulk" box. The tool pulls out every URL, turns the slug into a readable title, and adds it as a link. Tick "Group by top-level folder" and it will also sort /blog/ pages, /docs/ pages, and so on into their own sections.

Treat that as a starting point, not the finished file. Delete the pages that don't matter, tidy the titles, and write a real description for each link that survives. A sitemap with 800 URLs should probably become an llms.txt with 15 to 40.

No. The generator runs entirely in your browser and never sends what you type to our server or anyone else's. It does keep a copy in your browser's local storage so you don't lose your work if the page reloads; "Start over" wipes it. It also doesn't fetch your pages, so the built-in checks look at the file's format (a title, valid https links, no duplicates, no stray headings), not at whether each URL actually loads. Click through your links before you publish.