How to create an llms.txt for your Shopify store.

What llms.txt is

The llms.txt proposal, published in late 2024 and gaining traction through 2025, is a community convention for a plain-text Markdown file served at /llms.txt. It gives AI crawlers a human-readable map of a website: title, one-paragraph summary, then annotated links to the most important pages. It sits alongside robots.txt and sitemap.xml but serves a different purpose. Robots.txt says what crawlers can do; sitemap.xml lists every indexable URL; llms.txt summarizes the site and points at a curated set of important pages.

As of Q1 2026, llms.txt is not a W3C standard and no AI engine formally documents its behavior with the file. eCommerce Insights's crawler log observations across hundreds of Shopify stores show that GPTBot, ChatGPT-User, PerplexityBot, ClaudeBot, and Google-Extended do fetch /llms.txt on sites that publish one. Adoption is advisory, not binding. Publishing one costs an hour of work and gives AI crawlers a better read of the store.

Why Shopify merchants in particular need it

A typical Shopify store has a few hundred to a few thousand URLs. The homepage, twenty collections, a hundred PDPs, twenty blog posts, policy pages, and a long tail of filter-variant URLs that add little. A raw sitemap exposes all of them with no priority signal. An AI crawler that only fetches a handful of pages needs help picking the right ones. The llms.txt is that help.

On Shopify specifically, the default theme produces a sitemap with every product and collection, including drafts where they have been indexed in error. A curated llms.txt narrows that to the twenty to fifty pages that genuinely describe the brand and its flagship catalog. eCommerce Insights's own audit data (Q1 2026, illustrative) shows AI crawlers fetching deeper portions of the site after a good llms.txt is published, measured by unique URL fetches per bot per week.

A sitemap lists every door. An llms.txt says which rooms actually matter.

Step 1: Confirm which pages matter for AI crawlers

Start with a list. For a typical D2C Shopify store this is: the homepage, the about or brand-story page, the top ten collections by revenue, the top twenty PDPs (hero SKUs plus anchor products), the policy set (shipping, returns, sizing, sustainability), any guides or editorial that actually answers shopper questions, and a contact page. Skip cart, checkout, and account pages. Skip draft or archived products. Aim for 40 to 80 entries total.

Pull the list from Shopify admin by sorting collections and products by revenue over the last 90 days. Export to a spreadsheet, add a one-sentence description for each URL that tells an AI what the page is about, and you have the raw material for the file.

Step 2: Pick a location to serve llms.txt from (Shopify hosting quirks)

The convention says the file must live at yourdomain.com/llms.txt. Shopify's hosting, as of Q1 2026, does not let a merchant upload arbitrary files directly to the domain root. Three working approaches:

• Shopify Files plus URL redirect. Upload llms.txt to the Files section in Shopify admin (which gives it a public CDN URL like cdn.shopify.com/s/files/...), then add a URL redirect in Shopify admin from /llms.txt to that CDN URL. Simple, requires no infrastructure.
• Cloudflare or edge worker rewrite. If the storefront sits behind Cloudflare, a Worker or a Page Rule can rewrite requests for /llms.txt to serve the text content directly. Cleanest result; no redirect chain.
• Headless frontend. Hydrogen, Next Commerce, or a custom frontend can serve /llms.txt as a static file. The easiest path of all, but only relevant for headless stores.

Step 3: Draft the file (with template)

The llmstxt.org proposal specifies: an H1 title, a blockquote summary, then Markdown sections with annotated links. The format is intentionally Markdown so LLMs can consume it directly. Keep descriptions short. Keep sections logical: Product, Collections, Guides, Policies, Contact.

The llms.txt generator produces a first draft from a Shopify storefront URL in under a minute. Most merchants will still want to hand-edit the descriptions to match brand voice.

Step 4: Add structured links to collections and key PDPs

Every link in the file has three parts: the URL, a short label, and a one-sentence description that tells an AI crawler what the page offers. Avoid marketing copy. Write the description the way a Wikipedia editor would write a summary. For a PDP: "The Acme Signature Tee — 100% organic cotton, heavyweight 8oz fabric, unisex fit, available in four colors, $48." For a collection: "Men's outerwear — 24 SKUs, weather-resistant jackets and insulated parkas built for the Pacific Northwest."

Keep link lists under fifty per section. Past that, crawler attention drops. If a section needs more, create a second dedicated landing page and link to that instead.

Step 5: Host via Shopify's file assets or a redirect

In Shopify admin, open Settings → Files. Upload the llms.txt file. Copy the resulting CDN URL. Then open Online Store → Navigation → URL Redirects and add a new redirect: from /llms.txt to the CDN URL just copied. Save and test.

Caveat: some Shopify themes and apps intercept root-level paths. Test before assuming it works. A CDN-level rewrite avoids the redirect entirely and is the more durable path for stores that already use Cloudflare or a similar front-door provider.

Step 6: Validate the file

Open a private browser window and navigate to https://yourdomain.com/llms.txt. Check that the page loads without a 404, the content is plain text (not rendered as HTML wrapped in the Shopify theme), the Content-Type header is text/plain or text/markdown, and there is no long redirect chain. Click through five or ten of the listed URLs to confirm they all respond 200.

The llms.txt generator includes a built-in validator that runs all these checks in one go. eCommerce Insights's audit flags common errors: Content-Type set to text/html, the file returning a redirect loop, or the file accidentally being served behind the Shopify password page on development stores.

Step 7: Monitor whether AI engines actually fetch it

Publishing the file is half the job; the other half is knowing it gets read. Shopify does not expose raw server logs on standard plans, but Cloudflare, Fastly, and similar providers do. Filter for user agents GPTBot, ChatGPT-User, PerplexityBot, ClaudeBot, Google-Extended, Bingbot, and path /llms.txt. Log fetch frequency per bot per week.

As of Q1 2026, a healthy D2C Shopify store typically sees two to ten fetches of /llms.txt per week across the major AI bots (eCommerce Insights observation, illustrative). Zero fetches after four weeks means something is wrong — usually the redirect chain, the Content-Type, or the file being gated behind a theme password.

Common mistakes

• Hosting the file at /pages/llms-txt instead of /llms.txt. Crawlers will not find it.
• Wrapping the content in the theme's HTML shell. It must be plain text.
• Setting Content-Type to text/html. Use text/plain or text/markdown.
• Including draft, archived, or out-of-stock products in the links. Cite only pages shoppers can actually act on.
• Letting the file go stale after a collection relaunch. Regenerate monthly.
• Over-marketing the descriptions. AI crawlers reward specificity, not superlatives.

A dead link in llms.txt is worse than no llms.txt at all. Treat it like a production asset.

Code example: a working llms.txt for a Shopify apparel brand

Below is an illustrative sample for a fictional apparel brand called Ravenback. Copy, adapt, and replace the URLs and descriptions with the brand's real pages.

# Ravenback > Ravenback makes heavyweight, Pacific-Northwest-built outerwear and basics for everyday wear. Founded in Portland in 2018. Shopify-native D2C. Ships to the US, Canada, UK, EU, and Japan. ## Brand - [Homepage](https://ravenback.com/): Brand overview, featured collections, latest drops.
- [About Ravenback](https://ravenback.com/pages/about): Founders, manufacturing partners in Portland and Vancouver BC, materials sourcing.
- [Sustainability](https://ravenback.com/pages/sustainability): Organic cotton sourcing, recycled wool supply chain, annual impact report. ## Collections - [Men's outerwear](https://ravenback.com/collections/mens-outerwear): 24 SKUs — insulated parkas, waxed canvas jackets, puffer vests.
- [Women's outerwear](https://ravenback.com/collections/womens-outerwear): 22 SKUs — trench coats, field jackets, shell parkas.
- [Heavyweight tees](https://ravenback.com/collections/heavyweight-tees): 12 SKUs — 8oz unisex tees in organic cotton, across four color ways.
- [Wool knits](https://ravenback.com/collections/wool-knits): 18 SKUs — cardigans, pullovers, beanies in recycled merino.
- [Accessories](https://ravenback.com/collections/accessories): 30 SKUs — belts, caps, tote bags, leather wallets. ## Flagship products - [The Founder Parka](https://ravenback.com/products/founder-parka): 650-fill responsibly-sourced down, water-resistant shell, adjustable hood. Unisex, sizes XS-XXL, $485.
- [Signature Heavyweight Tee](https://ravenback.com/products/signature-heavyweight-tee): 8oz organic cotton, unisex fit, four colors (bone, slate, olive, black), $48.
- [Cascade Field Jacket](https://ravenback.com/products/cascade-field-jacket): Waxed canvas, four-pocket, made in Portland. Men's and women's cuts, $295.
- [Terminal Wool Cardigan](https://ravenback.com/products/terminal-wool-cardigan): Recycled merino, shawl collar, five buttons, hand-linked. Unisex, $245. ## Guides - [Sizing and fit](https://ravenback.com/pages/sizing): Men's and women's measurement charts, fabric stretch notes.
- [Care and repair](https://ravenback.com/pages/care-repair): Wool, cotton, and waxed-canvas washing and mending instructions.
- [Wool vs synthetic insulation](https://ravenback.com/blogs/field/wool-vs-synthetic): Category-level explainer, trade-offs by climate and use case. ## Policies - [Shipping](https://ravenback.com/policies/shipping-policy): Rates, transit times, international duties.
- [Returns and exchanges](https://ravenback.com/policies/refund-policy): 30-day returns, free US exchanges, international return rules.
- [Privacy](https://ravenback.com/policies/privacy-policy): Data handling, email list, cookie usage.
- [Terms of service](https://ravenback.com/policies/terms-of-service): Purchase terms, warranty language. ## Contact - [Contact page](https://ravenback.com/pages/contact): Email, support hours (Pacific), wholesale inquiries.
- [Stockists](https://ravenback.com/pages/stockists): Retail partners carrying the line in US, Canada, Japan.

Key takeaways

• Serve /llms.txt at the apex domain, not a subpath or page route.
• On Shopify, use a Files upload plus URL redirect, or a CDN-level rewrite.
• Curate 40 to 80 entries — top collections, flagship PDPs, guides, policies.
• Validate Content-Type, chase redirects, check every link responds 200.
• Monitor crawler logs for GPTBot, PerplexityBot, ClaudeBot weekly.

Related guides

Tools and product

• llms.txt generator — draft and validate an llms.txt from a Shopify storefront URL.
• The eCommerce Insights product — tracking, scoring, and PDP recommendations that pair with a well-formed llms.txt.