llms.txt Generator
Create an llms.txt file for your site - the emerging standard that tells AI assistants what your site is about and which pages matter most.
Upload to yoursite.com/llms.txt. Keep it short — it's a map for models, not a sitemap of everything.
Key takeaways
- ✓llms.txt is a markdown file at your domain root that describes your site and curates its most important pages for language models.
- ✓It is not robots.txt. robots.txt controls access; llms.txt provides context and curation.
- ✓Adoption is early and not universal — treat it as a cheap bet on an emerging standard, not a guaranteed ranking factor.
- ✓Curate ruthlessly. Ten well-chosen pages beat two hundred, because the file is a map, not a sitemap.
- ✓The exercise of writing it is valuable on its own: if you cannot describe your site in one sentence, models cannot either.
What llms.txt actually is
llms.txt is a plain markdown file that lives at the root of your domain, at yoursite.com/llms.txt. It contains a title, a one-line summary of what your site is, optional context, and a curated set of links grouped into sections.
The idea behind it is straightforward. Language models working with a website face a problem: your site might have thousands of URLs, most of them irrelevant to understanding what you do. Navigation menus, faceted filters, paginated archives and cookie banners all add noise. A model trying to understand your business has to infer it from whatever pages it happens to encounter.
llms.txt is an attempt to remove that guesswork. Instead of inferring, a model can read a short file where you state plainly what your site is and point to the handful of pages that best explain it.
How llms.txt differs from robots.txt and sitemap.xml
These three files are frequently confused, and they do genuinely different jobs.
- robots.txt — permission. Which crawlers may access which paths. It is a set of rules, not a description.
- sitemap.xml — inventory. Every URL you want indexed, machine-readable, exhaustive by design.
- llms.txt — orientation. What this site is, and which pages matter most, written for a reader that needs context rather than coverage.
A sitemap answers "what exists here?" llms.txt answers "what should I read to understand this?" The difference is curation. A sitemap listing five thousand URLs is doing its job correctly. An llms.txt listing five thousand URLs has failed completely.
Worth stating plainly: llms.txt is a proposed convention, not a ratified standard, and support across AI platforms is inconsistent. Nobody should promise you rankings from it. The argument for adding it is that it costs ten minutes and the downside is zero.
What to include in your llms.txt
The summary line
One sentence describing what your site is and who it serves. This is the hardest part of the file and the most valuable. If you cannot state it cleanly, that is a signal worth acting on — a model reading your homepage is facing the same difficulty.
Good: "Specialty coffee roaster shipping fresh beans across Europe, roasted to order." Weak: "Your partner in premium beverage experiences."
The pages worth listing
Include the pages that explain your business or answer real questions:
- Core product or service pages
- Pricing, if it is public
- Documentation or how-to guides
- Your about page, if it establishes genuine expertise
- Cornerstone articles that define your position on a topic
Leave out blog archives, tag pages, thin category listings, legal boilerplate and anything you would not send to a new customer as a starting point.
Section grouping
Group links under headings that describe their function — Products, Docs, Guides, Company. The grouping itself carries meaning: it tells a model how you think your own site is organised, which is often more informative than the individual links.
How to publish the file
The file must resolve at the root of your domain. On most platforms this means uploading it as a static file. On Shopify, where root-level file access is restricted, the usual approach is to create a page and set up a redirect, or serve it through your theme where your setup allows.
Once it is live, verify it loads as plain text in a browser at yoursite.com/llms.txt. If it renders inside your site template with headers and navigation around it, it is not serving correctly — the file should be raw markdown and nothing else.
Is it worth doing yet?
This deserves a straight answer rather than enthusiasm.
Support is genuinely partial. Some tools and platforms read llms.txt; many do not. There is no public evidence that any major assistant treats it as a strong signal, and anyone claiming otherwise is guessing.
The case for doing it anyway rests on asymmetry. The cost is one short file and ten minutes. The downside risk is zero — a file nobody reads harms nothing. The upside, if adoption grows, is that you were already there. That is a reasonable bet at that price.
The secondary benefit is more reliable: writing the file forces you to articulate what your site is and which pages matter. Most teams find that harder than expected, and the clarity that comes out of the exercise tends to improve homepage copy and internal linking regardless of what any model does with the file.
llms.txt in the wider AEO picture
Treat this as one small piece of a larger job. If AI assistants are not mentioning your brand, llms.txt is unlikely to be the reason, and it is unlikely to be the fix.
The higher-impact work sits elsewhere: confirming AI crawlers can actually reach your site, structuring pages so answers are extractable, adding schema so models can identify who you are, and building presence on the third-party sources models actually cite. llms.txt supports that work; it does not substitute for it.
The sensible sequence is to check crawler access first, fix page structure second, and add llms.txt as a finishing touch once the fundamentals are in place.
A worked example of a good llms.txt
Structure matters less than restraint, but a consistent shape helps. A well-formed file reads roughly like this:
- An H1 with the site or brand name.
- A blockquote summary — the one-sentence description of what the site is.
- An optional paragraph of context: founding, specialism, market, anything that disambiguates you from similarly named businesses.
- H2 sections grouping links by purpose, each link followed by a short description of what the page contains.
The link descriptions carry more weight than people expect. "Pricing" tells a model almost nothing. "Pricing — plan tiers, per-seat costs and what is included at each level" tells it what it will find before it decides whether to fetch the page.
Common mistakes
Treating it as a second sitemap
The single most frequent error. Exporting every URL into markdown defeats the purpose entirely. If a model has to read four hundred links to understand your business, you have recreated the problem the file was meant to solve.
Marketing language in the summary
The summary line is read by a machine trying to categorise you. Superlatives and positioning statements carry no information. Say what you sell, to whom, and where.
Letting it go stale
A file listing discontinued products and dead URLs is worse than no file, because it actively misinforms. Review it whenever your site structure changes meaningfully.
Serving it as HTML
If the file loads wrapped in your site template, it is not serving as plain text. Check it in a browser and confirm you see raw markdown with no navigation around it.
Final thoughts
llms.txt is a low-cost, low-risk, uncertain-reward addition to a site. That combination makes it easy to recommend and easy to over-hype, and plenty of people are currently doing the second.
Generate the file, publish it, move on to the work that reliably matters. The ten minutes you spend deciding which pages deserve a place on that list will teach you more about your own site architecture than the file itself is likely to earn you.
Frequently asked questions
What is llms.txt?
A plain markdown file at yoursite.com/llms.txt that describes what your site is and links to its most important pages, written specifically for AI models rather than crawlers or human visitors.
Do AI assistants actually read llms.txt?
Support is partial and inconsistent. Some tools and platforms read it; there is no public confirmation that major assistants weight it heavily. It is best understood as a low-cost bet on an emerging convention.
Is llms.txt the same as robots.txt?
No. robots.txt controls which crawlers can access which paths — it is about permission. llms.txt provides context and curation, describing what your site is and which pages matter.
Where do I upload llms.txt?
At the root of your domain, so it resolves at yoursite.com/llms.txt. It should serve as raw markdown, not rendered inside your site template.
How many pages should I list?
Usually between five and twenty. The file is a curated map, not a sitemap. If you are listing everything, you have misunderstood the purpose.
Will llms.txt improve my Google rankings?
There is no evidence it affects traditional rankings. Google's crawling and indexing rely on robots.txt, sitemaps and links. Treat llms.txt as an AI-visibility experiment, not an SEO tactic.
Do I need llms.txt if I already have a sitemap?
They serve different purposes. A sitemap lists everything for indexing; llms.txt curates the few pages that best explain your site. Having one does not replace the other.
Related free tools
The free mini SEO audit checks crawler access, page structure and entity clarity, then tells you what to fix first.