Skip to main content

Growth100X

AI Search & AEO10 min readUpdated July 2026
SBy  Sumit Sagar · Founder, Growth100X
The answer, straightTL;DR
what is llms.txt and do i need it?

llms.txt is a plain-text file at your site root — like robots.txt, but written for AI models instead of crawlers — that maps your most important pages in clean markdown. On its own it won’t rank you: a 2026 study of 137,000 sites found 97% of llms.txt files get zero AI requests. Paired with clean structure and schema, it’s still a five-minute addition worth having.

Somewhere between a client asking “should we have an llms.txt file?” and a developer shrugging “I mean, it can’t hurt,” most SMB owners never get a straight answer on what is llms.txt — what it does, whether it’s worth the hour it takes to build, and whether AI engines actually read it. Here’s the honest version, not the hype version.

28%
OF ANALYSED DOMAINS HAVE ONE
97%
OF THOSE FILES GET ZERO REQUESTS
2024
YEAR THE STANDARD WAS PROPOSED
In this guide
01 · The definition

What is llms.txt, and do you actually need one?

llms.txt is a plain-text file, written in markdown, that sits at your domain root — yoursite.com/llms.txt — and gives AI models a short, curated index of your most useful content. Proposed by developer Jeremy Howard in 2024 and documented at llmstxt.org, the format takes deliberate inspiration from robots.txt: a single, predictable file location that any crawler — human-built or AI-built — knows to check.

The spec itself is narrow by design. A valid llms.txt file has an H1 with your site or project name, a blockquote with a one-line summary, an optional paragraph of context, then one or more H2 sections listing markdown links to your key pages with a short note after each. No styling, no images, no fluff — just the pages you’d want an AI model to read first if it could only read five things.

Whether you need one depends on what you’re trying to solve. It is not a ranking factor, it does not replace clean on-page structure or schema markup, and it will not by itself get you cited by ChatGPT or Perplexity — that work is covered properly in how to get cited by ChatGPT and Perplexity. What llms.txt does is remove friction for any AI system that does try to parse your site, the same way a sitemap removes friction for Googlebot without being a ranking signal on its own.

02 · The honest section

Does llms.txt actually help you get cited — yet?

Here’s the part most llms.txt guides skip: the evidence that it currently influences AI citations is thin. An Ahrefs study of 137,000 domains found that while 28% of sites using Ahrefs’ own analytics had published an llms.txt file, 97% of those files were never fetched by anything at all in the month studied. Of the requests that did come through, only 19.5% came from named AI tools — and AI retrieval bots, the specific systems most likely to power a live citation, accounted for just 1.1% of total requests.

That doesn’t mean llms.txt is worthless — it means it isn’t doing the job most people assume it’s doing. Major AI providers have not confirmed that llms.txt influences retrieval or ranking inside their live answer engines, and separate adoption tracking puts it on only around 8.7% of top 1,000 domains as of mid-2026 — meaning most of the web still hasn’t bothered, and the sites racing to add one are often optimising for a signal nobody’s reading yet.

Where it plausibly matters is upstream, not at query time: as a clean, well-labelled source for any future training runs or ingestion pipelines, rather than something ChatGPT reads live before answering your prospect’s question. Treat it as cheap future-proofing, not a growth lever — the actual lever is still the content quality, structure and schema covered in our cross-engine GEO playbook.

“

llms.txt won’t make you rank. It just makes sure the answer engines that do bother to read your site don’t waste the visit.

File Audience Influences rankings?
robots.txt Search & AI crawlers Indirectly — controls access
sitemap.xml Search crawlers Indirectly — aids discovery
Schema markup Search & AI engines Yes — proven AEO/GEO signal
llms.txt AI models, agents Unconfirmed, low current uptake
03 · The format

A minimal llms.txt, line by line

Below is a complete, valid llms.txt file for a small services business — short enough to build in one sitting, structured exactly to spec.

/llms.txtExample
# Acme Design Studio > Boutique UI/UX and brand design studio for B2B SaaS companies, based in Austin, TX. Acme Design Studio helps SaaS teams ship interfaces that convert. Founded 2019, 40+ shipped products, remote-first team of 12. ## Docs - [Services overview](https://acmedesign.com/services/): what we do and who it's for
- [Case studies](https://acmedesign.com/work/): results across 12 SaaS clients
- [Pricing](https://acmedesign.com/pricing/): project and retainer rates ## Optional - [Blog](https://acmedesign.com/blog/): design and SaaS growth articles

Notice what’s missing: no navigation menu, no footer links, no cookie-banner copy, no marketing filler. That restraint is the entire point — an llms.txt file is only useful if it’s a genuine shortlist, not a re-export of your sitemap. Twenty links defeats the purpose as thoroughly as zero links does.

04 · The build

How to create and upload your llms.txt file

  1. Open a plain-text editor and create a new file named exactly llms.txt — no .html, no .md extension.
  2. Add an H1 with your business or site name, then a blockquote (>) with a one-sentence summary of what you do.
  3. Add one short context paragraph if needed — who you serve, where you operate, what makes you specific.
  4. Add H2 section headers (“Docs,” “Products,” “Optional”) each containing a short markdown-linked list of your 3–8 most important pages, with a one-line note after each.
  5. Upload the file to your site’s root directory so it resolves at yoursite.com/llms.txt — via FTP, your host’s file manager, or a plugin if you’re on WordPress.
  6. Re-check it every quarter as pages change; a stale llms.txt pointing at dead URLs is worse than none at all.

If you’re not comfortable editing server files directly, this is exactly the kind of half-hour technical task worth folding into a broader AI search and SEO engagement rather than handling piecemeal — it’s most valuable done alongside schema, sitemap and crawl-budget cleanup, not in isolation.

It’s also worth deciding upfront who owns the file long-term. On most SMB sites, llms.txt quietly rots within a year — a service gets discontinued, a pricing page moves, and nobody updates the one file that was supposed to be the “clean, curated” version of the site. Assign it to whoever already owns your sitemap or robots.txt, and put a quarterly reminder on the same calendar entry that triggers your other technical SEO checks, covered in more depth alongside measurement in how to track AI search traffic once the file is live and you want to know whether any of this is actually working.

Tip: check your server logs after 60–90 days for hits from GPTBot, PerplexityBot or ClaudeBot on /llms.txt. If you see genuine fetches, keep iterating the file. If you see none — which, per the data above, is the likely outcome for now — don’t over-invest further time in it.
05 · The bigger file

llms-full.txt, explained

Where llms.txt is an index, llms-full.txt is the whole book: a single flat markdown file containing the entire text of your key pages concatenated together, so an AI tool can ingest your full content in one fetch instead of following individual links. It’s most common on documentation-heavy sites — developer tools, SaaS platforms, technical products — where an AI coding assistant benefits from reading the whole reference at once.

For most SMB marketing sites, llms-full.txt is optional and often unnecessary — a services business rarely has the kind of dense reference content that benefits from a full-text dump, and maintaining a duplicate flat-file version of your whole site adds upkeep for a use case you probably don’t have yet. Start with llms.txt; add llms-full.txt only if you’re running documentation, a knowledge base, or a product with genuine technical depth.

If you do build one, treat it as a generated artifact rather than something hand-maintained page by page — export your core content to markdown, concatenate it, and regenerate the file whenever pricing, services or key pages change materially. A stale llms-full.txt that still describes a discontinued product or an old price is arguably worse than not having one, since it hands an AI system confidently wrong information dressed up as an authoritative source.

Want your whole site AI-readable, not just one file?

Growth100X implements llms.txt alongside full AI-readability work — schema, structure, crawl access — through our AI search & SEO service.

Book your free audit →

06 · Straight answers

FAQ

What is llms.txt and do I need it?

llms.txt is a plain-text, markdown-formatted file at your site root that gives AI models a curated index of your key pages. Most small businesses don’t strictly need one — it isn’t a ranking factor — but it takes under an hour to build and does no harm, so it’s worth having as part of a wider AI-readability setup.

Does llms.txt improve my Google rankings?

No. llms.txt has no confirmed effect on Google search rankings and is unrelated to traditional SEO signals. It’s built for AI models and agentic tools, not search engine crawlers, and shouldn’t be treated as an SEO tactic.

What’s the difference between llms.txt and llms-full.txt?

llms.txt is a short index of links to your most important pages. llms-full.txt is a single file containing the full text of those pages concatenated together, letting an AI tool ingest everything in one fetch. Most SMB sites only need the former.

Where do I upload llms.txt on my website?

It must sit at your domain’s root directory so it resolves at yoursite.com/llms.txt, the same way robots.txt does. Upload it via FTP, your hosting file manager, or an SEO plugin if you’re on WordPress.

Do ChatGPT, Perplexity and Gemini actually read llms.txt files?

Rarely, as of 2026. A large-scale study found AI retrieval bots accounted for only about 1.1% of requests made to llms.txt files, and no major provider has confirmed it directly powers live citations. It may matter more for future training data collection than for today’s answer engines.

S
Written by
Sumit Sagar — Founder, Growth100X

10+ years building growth systems for SaaS, fintech, healthcare and Web3. Ex-Head of Marketing at LCX — scaled 10K → 150K users and $50M+ raised across 12 token sales. Builds voice agents, automation and AI-search systems hands-on for SMBs.

Growth100X · Growth Systems

Want this built for your business?

We build AI growth systems for SMBs. Book a free 30-minute audit and we will map it to your funnel.

Explore Growth Systems →Book a free audit →