llms.txt is a simple, plain-text file at the root of your domain that gives AI language models a curated map of your site. Think of it as a sitemap written for answer engines instead of crawlers: it lists your most important pages with a one-line description of each, plus a short summary of what your store is, so a model can understand what you sell and which URLs to cite without crawling and interpreting your entire catalogue. This guide covers what belongs in the file, the exact format, and the real ways to serve it on Shopify — where the storefront root is not yours to write to.

Why it matters for a store

AI answer engines work with a limited context window and a strong preference for sources they can parse quickly. When a shopper asks ChatGPT or Perplexity "what is a good brand for X," the model is not going to render your JavaScript, wait for apps to load, and reverse-engineer your navigation. It reaches for the cleanest, most authoritative signal it can find. A well-written llms.txt is exactly that signal: you hand the model the highlights — key collections, best guides, policies, and the facts you want repeated — in the plain format it reads best.

The practical payoff is control. Without llms.txt, the model decides which of your pages matter and often guesses wrong, surfacing a stale blog post instead of your flagship collection. With it, you point directly at the pages that convert and the facts that are true. This is one layer of a broader answer engine optimization strategy, but it is the layer with the least effort and the clearest effect.

llms.txt is not robots.txt, and not a sitemap

These three files are easy to confuse, so keep the roles straight. robots.txt is a permissions file — it tells crawlers what they may and may not access. sitemap.xml is an exhaustive machine list of every URL on your site, built for search crawlers to discover pages. llms.txt is neither; it is an editorial highlights map written for language models, listing only the pages that matter and describing why each one is worth citing.

A useful way to hold it: robots.txt is the door policy, sitemap.xml is the full inventory, and llms.txt is the guided tour. You still want all three. The sitemap gets every page discovered; llms.txt makes sure the important ones get understood and quoted.

What a good llms.txt contains

The format is deliberately minimal — Markdown-style plain text, readable by a human and a model alike. A strong file for a Shopify store has four parts.

  • A single H1 with your store name, followed by a short blockquote or sentence summarising what you sell and what makes you different. This is the first thing the model reads, so make it factual and specific: "Handmade merino knitwear for cold-climate cyclists, shipped from Vermont."
  • A block of citable facts. Free shipping thresholds, return window, guarantee, materials, where you manufacture — the exact statements you want an AI to repeat back to a shopper. Write them as plain sentences a model can lift verbatim.
  • Grouped, linked lists of your most important pages under H2 headings like Collections, Guides, and Policies. Each entry is a Markdown link plus a one-line description of what the page covers and why it is worth citing.
  • Restraint. Twenty carefully chosen links beat two hundred. The point of the file is prioritisation, so a bloated llms.txt that mirrors your whole sitemap defeats itself.

The common mistake is treating it as another sitemap and dumping every URL in. The value is in the curation. If everything is important, nothing is.

How to add it on Shopify

Here is the Shopify-specific catch: Shopify owns and serves the storefront root, so you cannot drop a file at your-store.com/llms.txt through the admin the way you would upload an image, and on a .myshopify.com domain you have no root access at all. There are three reliable routes.

The first is an app proxy. A Shopify app can register a proxy path and serve the llms.txt content from its own backend under your storefront domain — this is how most SEO apps deliver it, because the file stays on your real domain and updates automatically. The second is a theme route: some merchants create a dedicated page and template and expose the content there, though matching the exact /llms.txt root path this way is limited by what Shopify routes to the theme. The third, for a store on a custom domain fronted by a CDN, is a lightweight edge worker (Cloudflare Workers and similar) that intercepts the /llms.txt request and returns your file before it ever reaches Shopify.

For most merchants the app-proxy route is the least fragile, because it keeps the file on-domain and removes the manual step. Whichever route you pick, confirm the file actually serves by opening your-store.com/llms.txt in a browser — a surprising number of setups return the storefront 404 page and no one notices. If you want the content generated for you from your live catalogue, the llms.txt generator builds a valid file structured the way models expect.

Keep it current, and pair it with crawler access

A stale llms.txt is worse than none, because it actively points AI at collections you have discontinued and facts that are no longer true. Regenerate it whenever you add or retire a collection, change shipping or return terms, or publish a major guide. Treating it as a set-and-forget file is the slow failure mode we see most.

llms.txt also does nothing on its own if the AI crawlers cannot reach your site in the first place. Make sure GPTBot, ClaudeBot, PerplexityBot, Google-Extended, and the rest are allowed in your rules — the Shopify robots.txt guide covers that, and the wider AEO for Shopify guide shows how the two fit together. Then verify you are actually being cited with an AI visibility check across the major engines.

RankEngine generates and maintains your llms.txt automatically, serves it on your domain, keeps it in sync as your catalogue changes, and rolls it into the same AI visibility scoring it uses for the rest of your answer-engine signals — so the file stays true instead of drifting out of date. See how it fits the full picture in the AEO tools overview.