What Is llms.txt? (And Should Your Site Have One)

What llms.txt is

llms.txt is a plain Markdown file, served at the root of your domain (yoursite.com/llms.txt), that gives AI systems a curated, human-written summary of what your site is and links to the pages that matter most. It was proposed in September 2024 by Jeremy Howard (Answer.AI) as a lightweight standard, explicitly modeled on robots.txt and sitemap.xml — a well-known convention, easy to adopt, easy to parse.

A minimal example:

# Peak Visibility

> AI SEO tool that generates sitemap.xml, schema markup, meta tags, and robots.txt for any website.

## Docs
- [Free SEO Checker](https://www.peakvisibility.io/free-seo-checker): Instant 10-point SEO check, no signup
- [How it works](https://www.peakvisibility.io/features): What gets generated and how

## Guides
- [Schema Markup Guide](https://www.peakvisibility.io/blog/schema-markup-guide): How to add structured data without coding
- [Sitemap Guide](https://www.peakvisibility.io/blog/how-to-generate-a-sitemap): How to generate and submit a sitemap

That's the whole format: a title, a one-line summary, and Markdown-style link lists grouped under headers. No schema to validate against, no required fields beyond the title — which is both the appeal (five minutes to write) and the limitation (nothing enforces that anyone actually reads it).

What llms.txt is not

It's easy to over-attribute what this file does. To be direct about the limits:

• It's not a ranking factor. No major AI system has confirmed that llms.txt affects whether or how often you're cited. It's a convenience file, not a scored signal.
• It's not universally read. Unlike robots.txt, which every major crawler respects as a directive, llms.txt adoption by AI systems is inconsistent and largely unconfirmed at the model-provider level. Some tools and plugins read it; there's no guarantee ChatGPT, Claude, or Perplexity's live retrieval systems parse it as of this writing.
• It doesn't replace good content structure. The things that actually drive AI citations — extractable answers, clear entity definitions, distributed brand mentions — matter far more than this file. See our GEO vs SEO breakdown for what actually moves the needle.

It's best understood as a hedge: low effort, plausible future upside, zero downside. Not a growth lever on its own.

Should your site have one?

Given the cost — a five-minute file with no maintenance burden — yes, there's little reason not to. It's most worth prioritizing if:

• Your site has a large or complex structure where an AI system's default crawl might miss your best pages
• You have documentation, an API, or reference content that benefits from being explicitly pointed to (llms.txt originated in developer-tool contexts for exactly this reason)
• You're already investing in GEO and want every plausible signal covered, even unconfirmed ones

It's lower priority if you haven't yet nailed the fundamentals — a real sitemap, working meta tags, actual schema markup. Those have confirmed impact. llms.txt doesn't, yet.

How to write one

1. Start with an H1 of your site/product name
2. Add a one-line blockquote summary of what the site does
3. Group your most important pages under H2 headers (Docs, Guides, Product, etc.) as Markdown links, each with a short description
4. Keep it to your actual highest-value pages — a few dozen links, not your entire sitemap. If it lists everything, it tells an AI system nothing about what matters most.
5. Save it as llms.txt in your site's root directory (not a subfolder) so it's reachable at yoursite.com/llms.txt

Some sites also publish an llms-full.txt — the same idea but with full page content inlined rather than just links, intended for cases where a crawler wants the content directly rather than following each link. That's a heavier lift and worth doing only once the basic file is in place and you have a specific reason (e.g., documentation sites where inlined content genuinely helps).

Check whether you already have one

Peak Visibility's free AI crawler checker confirms whether llms.txt exists on your domain, alongside checking whether your robots.txt is accidentally blocking GPTBot, ClaudeBot, or PerplexityBot — the more consequential of the two checks, since a robots.txt block has confirmed effects on AI citation, unlike llms.txt's unconfirmed ones.

FAQ

Does llms.txt improve my ranking in ChatGPT or Perplexity?

There's no confirmed evidence it does. It may help specific tools or plugins that explicitly parse it, but no major AI provider has stated their core retrieval system uses it as a ranking input. Treat it as a low-cost hedge, not a growth strategy.

Is llms.txt the same as robots.txt?

No. robots.txt is a directive — it tells crawlers what they're allowed to access, and major crawlers are built to respect it. llms.txt is informational — it's a suggestion of what matters, with no enforcement mechanism and inconsistent adoption.

Where do I put the file?

At your domain root: yoursite.com/llms.txt, exactly like robots.txt and sitemap.xml. A file in a subfolder won't be found by anything looking for it at the conventional location.

Do I need to update it often?

Update it when your most important pages change significantly — a major new product, a restructured docs site. It's not something that needs weekly maintenance like a sitemap does.