← Back to Blog

llms.txt — The New robots.txt for AI Crawlers (Complete Ecommerce Guide for 2026)

If you remember adding a robots.txt file to your website for the first time, you understand the moment we are in with llms.txt. A new standard is emerging — small in size, simple in concept, and quietly determining which sites AI assistants will surface and which they will skip.

For ecommerce stores, the stakes are direct. AI shopping traffic is small today (~1% of total) but growing 165 times faster than organic search. The merchants who add a well-structured llms.txt now will be cited by ChatGPT, Perplexity, and Google AI tomorrow. The merchants who do not will compete on increasingly thin organic-search margins.

This guide covers what llms.txt is, how it works, how to write one for an ecommerce store, and where major platforms stand on supporting it.

What Is llms.txt?

llms.txt is a plain markdown file placed at the root of your website (yoursite.com/llms.txt) that provides a structured, machine-readable overview of your site for large language models.

The proposal originated from Jeremy Howard at fast.ai in September 2024. The premise was simple: large language models have a context-window problem. They cannot crawl an entire site every time someone asks a question about it. They need a curated, concise summary — written by you — that tells them what your site is about and where the important information lives.

llms.txt fills that gap. It is the AI-era equivalent of a well-written README file for your entire website.

How llms.txt Differs From robots.txt and sitemap.xml

The three files serve distinct but complementary purposes:

FileAudiencePurposeFormat
robots.txtAll web crawlersWhat to crawl, what to avoidPlain text directives
sitemap.xmlSearch engine crawlersComplete URL inventoryXML
llms.txtAI models / LLMsWhat the site is about + curated key URLsMarkdown

You should have all three. They do not replace each other.

Why Ecommerce Sites Should Care

An AI assistant being asked "what specialty olive oil brands are direct-to-consumer?" needs to quickly understand what your store sells without crawling 5,000 product pages. A well-structured llms.txt gives that context in 200 lines or less.

Three reasons llms.txt matters more for ecommerce than for content sites:

  1. Product catalogs are large. AI models cannot ingest 5,000 product pages on demand. llms.txt provides the curated subset that matters.
  2. Buying intent is specific. When someone asks AI to recommend a product, the AI needs to know your store's specialty fast. A 100-word llms.txt summary delivers that.
  3. Competitive advantage is open. Almost no competitors have llms.txt yet. Adding one is a fast, low-cost differentiation move.

The llms.txt Spec

The format is intentionally simple. It is plain CommonMark markdown with a loose structure:

  1. H1 heading — the name of your site or brand
  2. Blockquote — a short summary of what the site does
  3. Optional paragraphs — additional context
  4. H2 sections — categorized lists of links to important pages

The links inside H2 sections are the meat of the file. Each link should be in markdown format, optionally followed by a short description of what the linked page contains.

Example llms.txt for an Ecommerce Store

Here is a complete example for a hypothetical specialty grocery store:

# Pantry & Co

> Independent specialty grocer in Portland, Oregon. Online ordering for ethnic foods, organic produce, butcher cuts, and natural wines. Pickup and local delivery.

Pantry & Co has been a family-run specialty grocer in Portland since 1998. We carry hard-to-find ethnic ingredients (Asian, Latin, Middle Eastern), USDA-certified organic produce, dry-aged meats from local farms, and natural wines from small producers.

We deliver within a 10-mile radius of Portland and offer pickup at our two locations. Online orders accepted 7 days a week.

## Categories

- [Asian Pantry](https://pantryandco.com/category/asian) — Soy sauces, miso, dashi, rice, noodles, snacks
- [Latin Pantry](https://pantryandco.com/category/latin) — Mole pastes, masa, dried chiles, tomatillos
- [Middle Eastern](https://pantryandco.com/category/middle-eastern) — Tahini, za'atar, sumac, preserved lemons
- [Organic Produce](https://pantryandco.com/category/produce) — Local seasonal organic produce
- [Butcher](https://pantryandco.com/category/butcher) — Dry-aged steaks, heritage pork, free-range poultry
- [Natural Wine](https://pantryandco.com/category/wine) — Small-producer natural wines, US delivery

## Featured Products

- [Spanish Saffron 1g](https://pantryandco.com/product/spanish-saffron) — Premium Mancha saffron, certified Coupe
- [Single-Origin Mole Negro](https://pantryandco.com/product/mole-negro) — Hand-prepared, Oaxaca-style
- [Maldon Sea Salt 8oz](https://pantryandco.com/product/maldon) — Pyramid flake sea salt

## Locations

- [Hawthorne Store](https://pantryandco.com/locations/hawthorne) — 3245 SE Hawthorne Blvd, Portland, OR
- [Mississippi Store](https://pantryandco.com/locations/mississippi) — 4015 N Mississippi Ave, Portland, OR

## Service

- [Delivery Info](https://pantryandco.com/delivery) — Local delivery within 10 miles
- [Pickup Info](https://pantryandco.com/pickup) — Same-day pickup at both locations
- [Wholesale Inquiries](https://pantryandco.com/wholesale) — For restaurants and cafes
- [Contact](https://pantryandco.com/contact) — Phone, email, hours

That entire file is under 1,500 characters. An AI model can ingest it in milliseconds and use it to surface Pantry & Co when shoppers ask "where can I buy authentic mole negro online?"

What AI Crawlers Do With llms.txt

Different AI engines treat llms.txt with different levels of priority. As of 2026:

  • ChatGPT (with browsing) — checks llms.txt when crawling a site for the first time. Uses it to understand site structure.
  • Perplexity — actively uses llms.txt in its retrieval pipeline. Sites with well-structured llms.txt are more likely to be cited.
  • Anthropic Claude (with web tools) — uses llms.txt when available. Documented in Claude's web-fetching behavior.
  • Google AI Overviews — does not officially endorse llms.txt yet but anecdotal evidence suggests it influences rankings.
  • Microsoft Copilot — uses Bing's index; llms.txt is processed through Bing's general web crawling.

The pattern is clear: llms.txt is being adopted by AI vendors faster than any web standard in the past decade. The momentum is real.

How to Add llms.txt to Your Site

Two approaches: manual or automatic.

Manual: write and update by hand

Create a plain text file named llms.txt with the markdown content described above. Upload it to the root of your domain. Test by visiting yoursite.com/llms.txt in a browser — it should render as plain text.

The downside: every time you add or remove a product, change a category, or update a featured item, you need to remember to update the file. Most merchants forget within 30 days.

Automatic: use a platform that generates it

The cleaner approach is to use an ecommerce platform that auto-generates llms.txt from your product catalog. The platform should regenerate the file every time you add, edit, or delete a product, so the file is always current.

BusinessCart.ai generates llms.txt automatically for every storefront. The file updates whenever your catalog changes — no manual maintenance required.

Platform Support: Where Each Major Platform Stands

Platformllms.txt supportNotes
ShopifyManual onlyUpload via theme code injection; no auto-update on catalog changes
Shopify PlusManual onlySame limitation despite higher tier
WooCommerceManual or third-party pluginA few SEO plugins now offer llms.txt generation
BigCommerceManual onlyNo native support as of 2026
SquarespaceNot possibleTemplates lock root file access
WixNot possibleSame limitation as Squarespace
WebflowManual via custom codePossible but requires technical setup
BusinessCart.aiAutomatic, always currentGenerated from product catalog on every change

Best Practices

Keep it under 200 lines

llms.txt should be a curated overview, not an exhaustive index. Use sitemap.xml for the full URL inventory. Reserve llms.txt for the URLs that matter most — categories, featured products, key informational pages.

Lead with specificity

The blockquote summary at the top is the most important line in the file. Be specific: "Pantry & Co is an independent specialty grocer in Portland, Oregon, with online ordering for ethnic foods, organic produce, butcher cuts, and natural wines" beats "Pantry & Co sells groceries online."

Update with every catalog change

An llms.txt that lists products you no longer carry actively hurts you — AI models will surface dead links, and your authority drops. Use a platform that updates llms.txt automatically, or commit to a weekly manual review.

Include geographic and audience specificity

If you serve a specific geography, market, or buyer type, say so. AI models filter on this. "Independent specialty grocer in Portland, Oregon" is dramatically more useful than "specialty grocer."

The Future of the Standard

llms.txt is not yet a formal W3C standard. There is no official body governing the spec. As of 2026, it is a community convention with strong adoption among AI vendors and tech-forward sites.

The risk of waiting is real. Every quarter you do not have llms.txt, AI models are training and crawling without your context. Your competitors who add it now compound their citation advantage.

The risk of adopting is minimal. Even if the standard evolves, the principles will not change: AI models need a curated, concise overview of your site. Whatever the file is called next year, the content you write today will translate.

Bottom Line

llms.txt is small, simple, and disproportionately impactful. It takes 30 minutes to write the first version and a few minutes to update going forward. The merchants who add it now will be cited by AI engines with increasing frequency through 2026 and beyond.

If your platform does not support llms.txt natively, you have two options: write and maintain it manually, or move to a platform that handles it for you.

See how BusinessCart.ai auto-generates llms.txt for every storefront →

Related: How to Get Your Products Cited by ChatGPT (Step-by-Step Guide for 2026)