llms.txt for Shopify: the working walkthrough

What llms.txt is (and is not)

The llms.txt proposal, published in late 2024, is a community convention for a Markdown file served at /llms.txt: an H1 title, a blockquote summary, then sections of annotated links to the site's most important pages. It complements rather than replaces the existing files — robots.txt says what crawlers may do, sitemap.xml lists every URL, and llms.txt says which pages actually matter and why.

Honest status, as of mid-2026: it is not a W3C standard, and no AI engine formally documents reading it. Crawler-log observations across Shopify stores show GPTBot, ChatGPT-User, PerplexityBot, ClaudeBot, and Google-Extended fetching the file where one exists. Adoption is advisory — but it costs an hour and gives crawlers with limited fetch budgets a better read of the store.

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

Why Shopify stores in particular need it

A typical Shopify store exposes hundreds to thousands of URLs — every product, collection, blog post, plus a long tail of filter variants that add nothing. The default sitemap presents all of them with no priority signal. An AI crawler that fetches a handful of pages per visit needs help choosing; a curated llms.txt narrows the field to the 40–80 pages that genuinely describe the brand and its flagship catalog. The same curation logic feeds the rest of the AI-readability stack covered in AI SEO for ecommerce.

Step 1 — List the pages that matter

For a typical D2C store: the homepage, the brand-story page, the top ten collections by 90-day revenue, the top twenty PDPs (hero SKUs plus anchors), the policy set (shipping, returns, sizing), genuinely useful guides, and a contact page. Skip cart, checkout, account pages, and draft or archived products. Pull the revenue ranking from Shopify admin, export to a sheet, and write a one-sentence description per URL.

Step 2 — Draft the file

Write descriptions the way a neutral editor would, not the way the homepage hero does. For a PDP: material, weight, fit, price. For a collection: what it contains and for whom. A trimmed apparel-brand example:

Copy# Ravenback Supply Co.

> Pacific Northwest outerwear and base layers for year-round field use.
> 120 SKUs across parkas, fleece, and merino. Made-to-last construction,
> lifetime repair program, ships from Portland, OR.

## Products
- [Founder Parka](https://ravenback.example/products/founder-parka): Men's
  waterproof 3L shell parka, 650-fill recycled down, slate/olive/black, $420.
- [Merino 200 Base Layer](https://ravenback.example/products/merino-200-crew):
  200gsm merino crew, unisex XS–XXL, four colorways, $95.

## Collections
- [Men's outerwear](https://ravenback.example/collections/mens-outerwear):
  24 SKUs — weather-resistant jackets and insulated parkas.

## Policies
- [Shipping](https://ravenback.example/policies/shipping-policy): Free US
  shipping over $75; international to 40 countries.
- [Returns](https://ravenback.example/policies/refund-policy): 60-day returns,
  free exchanges, lifetime repair program.

Keep each section under fifty links — crawler attention drops past that. The free llms.txt generator drafts this structure from a storefront URL in under a minute; hand-edit the descriptions afterward.

Step 3 — Pick a hosting approach

Approach	How it works	Fit
Files + URL redirect	Upload to Shopify Files (CDN URL), redirect /llms.txt to it	most stores
Edge worker rewrite	Cloudflare Worker serves the content directly at /llms.txt	cleanest
Headless frontend	Hydrogen / Next serves it as a static file	headless only

The redirect approach is a 302 hop to the CDN rather than a file served at the root — observed crawler behavior follows it, but an edge rewrite avoids the chain entirely and is the more durable path for stores already behind Cloudflare. Headless brands have it easiest; see the headless solution page for the wider audit workflow.

Step 4 — Upload and redirect

1. Shopify admin → Settings → Files → upload llms.txt. Copy the resulting CDN URL.
2. Online Store → Navigation → URL Redirects → add a redirect from /llms.txt to the CDN URL.
3. Save, then test immediately — some themes and apps intercept root-level paths.

Step 5 — Validate

Open yourdomain.com/llms.txt in a private window and confirm it resolves to the current file, not a 404 or a stale upload. Check the response is text/plain or served as readable text rather than forcing a download. Then re-run your store through the agentic readiness grader — llms.txt presence is one of the checks, alongside the robots.txt and schema signals that make up the agent-readability score.

Step 6 — Maintain and monitor

Regenerate after collection restructures or major catalog shifts — monthly is a sensible default. Watch server or CDN logs for fetches by GPTBot, PerplexityBot, and ClaudeBot to confirm pickup; eCommerce Insights' Agent Lens does this per-bot automatically. And remember the file is one signal among several: the engines still weight the PDP itself far more heavily, so pair this with the schema work in schema for AI search and the measurement layer in the product AI visibility pillar.