Generate an llms.txt file — the emerging standard that tells AI assistants like ChatGPT, Perplexity and Gemini what your site is about and which pages matter. Built from your homepage; download and upload to your root.
⚡ Interactive demo — sample data
Generated an llms.txt with a site name, summary and key pages for this sample site. Upload it to your domain root at /llms.txt.
Site name set as the Markdown H1 from the page titleLooks good
Summary blockquote built from the page's meta descriptionLooks good
Key pages: 8 descriptive internal links collectedLooks good
Valid Markdown — H1, blockquote and bulleted linksLooks good
Upload to yourdomain.com/llms.txt (domain root) — not a subfolderWarning
Generate an llms.txt file — the emerging standard that tells AI assistants like ChatGPT, Perplexity and Gemini what your site is about and which pages matter. Built from your homepage; download and upload to your root.
How it works
Enter your website URL
Paste your homepage URL and run the generator. We fetch the page and read its title, meta description and internal links to build the starting point for your llms.txt — no manual copying of your site structure.
Review the generated llms.txt
You get a clean Markdown file with an H1 site name, a blockquote summary and a 'Key pages' list of your most relevant internal links — the exact structure the llms.txt convention expects an AI assistant to read.
Download and upload it to your domain root
Download the file, refine the summary and link list if needed, then upload it so it's reachable at yourdomain.com/llms.txt. AI assistants and crawlers that support the convention look for it there.
What we check
Site name (H1) — Pulls your page title into a top-level Markdown H1, the single required element of the llms.txt format. It's the name an assistant uses to identify the site the file describes.
Summary blockquote — Uses your meta description as the blockquote summary directly under the H1 — the short statement of what the site is and does that the convention places first for quick context.
Key pages list — Collects your most relevant internal links into a '## Key pages' Markdown list with descriptive anchor text, so assistants know which URLs matter most rather than crawling everything blindly.
Markdown formatting — Outputs valid Markdown — H1, blockquote and bulleted links — because the llms.txt convention is deliberately plain Markdown that's easy for both humans and language models to parse.
Internal links only — Filters to links on your own domain with meaningful anchor text, so the file describes your site rather than every outbound link, navigation icon or social button on the page.
Root-path readiness — Produces a single self-contained file meant to live at /llms.txt at your domain root — the canonical location the convention specifies, mirroring how robots.txt is discovered.
Common issues we catch
No meta description to summarize — If your homepage lacks a meta description, the summary blockquote comes out empty. Add a concise one-line description as your meta description (good for SEO too), or write the summary directly into the generated file.
Weak or generic anchor text — Links labeled 'Click here' or 'Read more' make a useless 'Key pages' entry because they don't say what the page is. Use descriptive link text on your site, or rewrite the labels in the generated file to name each page clearly.
Uploading to the wrong location — llms.txt must be reachable at the domain root — yourdomain.com/llms.txt — not in a subfolder. A file at /docs/llms.txt or /assets/llms.txt won't be found by tools that follow the convention.
Treating it like robots.txt — llms.txt isn't an access-control or blocking file — it doesn't allow or disallow crawling. It's a positive summary that helps assistants understand and cite your site. Use robots.txt for what you want blocked.
Listing too many or irrelevant pages — Dumping every URL defeats the purpose. The value is curation — point to your most important, content-rich pages. Trim the generated list to the handful that best represent what you want assistants to reference.
Letting it go stale — Like a sitemap, an llms.txt that lists deleted pages or an outdated summary misleads assistants. Revisit it when your site structure or positioning changes so the links and summary stay accurate.
Expecting guaranteed citations — llms.txt is an emerging, voluntary convention — not every assistant reads it, and having one doesn't guarantee you'll be cited. It improves the odds that supporting tools understand your site correctly; it isn't a ranking lever.
Where this matters
ChatGPT & other LLM assistants — The convention is aimed at large language model assistants that may fetch a site's llms.txt to get a curated summary and key links instead of parsing full HTML. Adoption is growing across AI tools.
Perplexity & answer engines — Answer engines that cite sources benefit from a concise, structured description of your site. A clean llms.txt makes it easier for them to identify what you offer and which page to reference.
Gemini & Google AI surfaces — As AI answers become a primary search surface, a clear machine-readable summary helps these systems attribute information to your brand and link to the right pages.
Documentation & SaaS sites — Docs-heavy and product sites were early adopters because llms.txt lets them point assistants straight at API references, guides and key product pages rather than marketing chrome.
Any website root — Because it's a single static Markdown file at /llms.txt, any site — WordPress, Shopify, custom, static — can adopt it by uploading one file, the same way you'd add robots.txt or a sitemap.
Frequently asked questions
What is llms.txt?
llms.txt is an emerging convention: a single Markdown file at your domain root that gives AI assistants a concise, curated summary of your site and a list of its most important pages. The idea is to help language models understand and reference your content accurately instead of guessing from raw HTML.
Where does the file go?
At your domain root, reachable as yourdomain.com/llms.txt — the same discovery pattern as robots.txt. It must be at the root, not in a subfolder, for tools that follow the convention to find it. Upload it however you'd upload any static file to your site.
Is llms.txt the same as robots.txt?
No. robots.txt controls which crawlers may access which paths — it's about blocking and allowing. llms.txt does the opposite: it's a positive, curated summary that helps AI assistants understand your site. They serve different purposes, and you can have both.
What format does llms.txt use?
Plain Markdown. The format starts with an H1 for the site name, a blockquote summary right below it, and then sections of links and notes. It's intentionally simple so both people and language models can read it without special parsing.
Will adding llms.txt help me get cited by AI?
It can improve the odds that assistants which support the convention understand your site and reference the right pages, but it's voluntary and not universally adopted, so it's not a guarantee. Think of it as making your site easy for AI to interpret correctly — a low-cost, forward-looking step.
What should the summary say?
One or two clear sentences on what your site is, who it's for and what it offers — the same clarity you'd want in a good meta description. The generator prefills it from your meta description; refine it so it reads as a confident, standalone description of your brand.
How many pages should I list?
Just your most important, content-rich pages — the ones you'd want an assistant to reference. Curation is the point, so a focused list of key pages beats dumping your whole sitemap. Trim the generated list down to what truly represents your site.
Do I need to update it over time?
Yes, when your site changes meaningfully. If you remove pages, restructure, or shift your positioning, refresh the link list and summary so assistants aren't working from stale information. It's a low-maintenance file, but an outdated one can mislead.
This is one of several free SEO tools from Custom Web Audits.
For a complete, prioritized analysis of your whole website,
run a full audit.
