# llms.txt for online stores: what to put in it

What llms.txt is, who says they read it, what a store’s file should hold and how to publish one on Shopify, WooCommerce, BigCommerce and Magento, with a builder that writes yours.

By Valerii Checha, 3 Nov 2026

What should an online store put in its llms.txt? Its name as a heading, one line on what it sells, where it ships and in which currency, then short lists of links to its main collections and to its shipping, returns and contact pages, with anything skippable under "Optional". llms.txt is a proposal, Google says its Search ignores the file, and we found no AI vendor that says its assistants read a store's copy. It takes half an hour to write and costs nothing to serve, so we would publish one anyway and keep expectations modest.

This guide explains what the file is, who says what about it, what a store's version should hold, and how to publish it on the main platforms, with a builder that writes one for you. Storewalk is ours: we build the scan, and its [llms.txt check](https://storewalk.ai/checks/DISC-002) is one of ours. The spec and the vendors are quoted from their own pages, read on 3 November 2026. The store mapping, which sections a shop's file should have, is ours.

## What is llms.txt?

[llms.txt](https://storewalk.ai/blog/agentic-commerce-glossary#llms-txt) is a Markdown file at `/llms.txt` that gives a language model a short map of a site. It is a proposal by Jeremy Howard, titled "The /llms.txt file, v2", first published in September 2024 and last changed in August 2026. No standards body has adopted it. The idea, in its own words: "We propose adding a /llms.txt markdown file to websites to provide LLM-friendly content".

The format is fixed in its order. First "An H1 with the name of the project or site. This is the only required section". Then "A blockquote with a short summary", then optional paragraphs with no headings, then sections under H2 headings that each list links as `- [name](url): notes`. A section called "Optional" is, by convention, for "links an agent can skip when a shorter context is needed."

Version 2 added two things. A file can now sit lower down: "The file can be placed at the site root, or at any path within it, covering the pages under that path", and "where more than one file applies, agents should use the most specific one." And a page can point to its file with a link marked `rel="describedby"`, beside the `rel="alternate" type="text/markdown"` link that points to a Markdown version of the page. The spec explains why the file does not live under `/.well-known/` with the other machine files: "well-known URIs exist only at the origin root, and many authors control only a path on a shared host". [How agents read a page](https://storewalk.ai/blog/how-agents-read-a-page) covers the Markdown side.

## Who reads llms.txt?

Each party below, in its own words:

Who says what about llms.txt, in their own words.

- Google Search: Says it ignores llms.txt and other AI text files, and that adding one neither helps nor harms. In its own words: “Doing so will neither harm nor help your site’s visibility or rankings in Google Search, as Google Search ignores them.” Source: [Google Search Central: AI optimization guide](https://developers.google.com/search/docs/fundamentals/ai-optimization-guide).
- Chrome Lighthouse: Its Agentic Browsing category, labelled experimental, checks that an llms.txt is present. Presence is what it describes. In its own words: “Checks for the presence of a machine-readable summary at the domain root.” Source: [Chrome for Developers: Agentic Browsing scoring](https://developer.chrome.com/docs/lighthouse/agentic-browsing/scoring).
- OpenAI, Anthropic, Gemini: The spec says the labs publish llms.txt files for their own developer docs. That is publishing; their crawler pages we read do not say their bots read yours. In its own words: “The AI labs themselves publish llms.txt files for their own developer docs: OpenAI, Anthropic, and Gemini.” Source: [llmstxt.org: The /llms.txt file, v2](https://llmstxt.org/).
- Shopify: Writes a default for every store and serves it at /llms.txt as well as /agents.md. In its own words: “The paths /llms.txt and /llms-full.txt also point to this content by default.” Source: [Shopify.dev changelog: customize llms.txt, llms-full.txt and agents.md](https://shopify.dev/changelog/posts/customize-llmstxt-llms-fulltxt-and-agentsmd).
- Storewalk (ours): Our DISC-002 is a small, optional check: it passes a file with an H1, at least 20 lines and links that resolve. In its own words: “llms.txt present and non-trivial” Source: [Storewalk: DISC-002](https://storewalk.ai/checks/DISC-002).

We found no statement from an AI vendor that its assistants or crawlers read a store’s llms.txt. That is an absence in the pages we read on 3 November 2026, not a denial.

Google is the clearest. Its AI optimization guide says "You don't need to create new machine readable files, AI text files, markup, or Markdown to appear in Google Search", and that adding them "will neither harm nor help". Chrome's Lighthouse goes the other way: its Agentic Browsing category, which Google calls "experimental and based on proposed standards" and which needs Chrome 150 or later, "Checks for the presence of a machine-readable summary at the domain root."

The AI labs publish llms.txt files for their own developer docs, and the spec names OpenAI, Anthropic and Gemini. OpenAI's, Perplexity's and Meta's crawler pages each link to an llms.txt, but it is the index of their own documentation. Publishing a file is not the same as reading yours. We read the crawler pages of OpenAI, Anthropic, Perplexity, Google, Apple, Meta and Amazon and found no statement that their bots read a site's llms.txt. None of them says they don't, either, and one blog post could change that.

It is still worth the half hour. Some tools, Lighthouse among them, look for the file, and writing one makes you list in plain words what you sell and where your policies are.

## What should a store put in its llms.txt?

The spec was written for software documentation, so a store has to translate it. Our mapping:

- The H1 is your store's name, exactly as customers know it.
- The blockquote says what you sell, where you ship and in which currency, in one or two sentences. In our view it is the line a model is most likely to repeat, so write a fact there and save the slogan for your homepage.
- A short paragraph can add the things a shopper's assistant would want to know first, such as a free returns window or a delivery promise. Only write what your policy pages say too.
- An H2 for products lists your main collections and a few best sellers, each with a few words of note.
- An H2 for policies and help lists shipping, returns, sizing and contact pages. We give them their own section because they answer a shopper's questions about delivery and returns.
- "Optional" holds the rest: about, blog, press.

Leave out cart, checkout and account URLs, which are no use to a model reading for context, and anything you would not want quoted. Link to Markdown versions of pages only if your site actually serves them; otherwise link to the normal pages.

Fill in the builder with your own store. It starts with an invented one, Northwind Linen, which ships linen to the EU and the UK and has never sold a thing:

An llms.txt builder for a store.

On the page, you fill in a store name, a one-line summary, a few more words, and links one per line as Title | URL | note, in three groups: products, policies and help, and optional. For an invented store it writes:

```markdown
# Northwind Linen

> Linen bedding and clothing, made in Portugal. We ship to the EU and the UK; prices are in EUR.

Every product page states its price, stock and delivery time. Returns are free within 30 days.

## Products

- [Bedding](https://example.com/collections/bedding): sheets, duvet covers and pillowcases
- [Shirts](https://example.com/collections/shirts): men’s and women’s, sizes XS to XXL
- [Dresses](https://example.com/collections/dresses)
- [Towels](https://example.com/collections/towels)
- [Gift cards](https://example.com/products/gift-card): from 25 EUR
- [New in](https://example.com/collections/new)
- [Sale](https://example.com/collections/sale)

## Policies and help

- [Shipping](https://example.com/policies/shipping-policy): costs and delivery times by country
- [Returns](https://example.com/policies/refund-policy): 30 days from delivery
- [Size guide](https://example.com/pages/size-guide)
- [Care guide](https://example.com/pages/linen-care)
- [Contact](https://example.com/pages/contact): email and phone, weekdays 9 to 17 CET

## Optional

- [About us](https://example.com/pages/about)
- [Journal](https://example.com/blogs/journal)
```

- Yes: Starts with an H1, the store’s name (the only part the spec requires) (the spec).
- Yes: Has a one-line blockquote summary (the spec).
- Yes: Lists at least one page as a Markdown link (our DISC-002).
- Yes: Has at least 20 non-empty lines (now 20) (our DISC-002).

Rather not type it all? The [llms.txt generator](https://storewalk.ai/tools/llms-txt-builder) is the same builder on a page of its own, with a button that drafts the file from your store's own pages.

## How does our scan check llms.txt?

Our [DISC-002](https://storewalk.ai/checks/DISC-002) asks for `/llms.txt` and reads it in a few steps. A missing file, or one that comes back as an HTML page, fails. The second happens on sites that answer every unknown path with their home page. A file passes when it has an H1, at least 20 non-empty lines and at least one link, and when the first five links load (we fetch only those on your own domain). Anything less is a WARN. It is one of the lightly weighted checks in our score.

The 20 lines are our threshold, not the spec's. A file with a name and a summary meets the spec and tells an agent almost nothing about a shop.

## How do you publish llms.txt on Shopify, WooCommerce, BigCommerce and Magento?

On Shopify you already have one. Shopify's docs say "Shopify manages an agents.md file by default for every store", and its changelog adds that "The paths /llms.txt and /llms-full.txt also point to this content by default." Shopify describes the data behind it as "auto-populated UCP and agent-interaction metadata for the store", which includes the store's UCP discovery and MCP endpoint URLs. To write your own, add a template to your theme under Online Store > Themes > Edit code. The three templates fall back on each other, which is easier to see than to read:

Shopify: which file each path serves.

| Templates in your theme | /agents.md                  | /llms.txt                   | /llms-full.txt              |
| ----------------------- | --------------------------- | --------------------------- | --------------------------- |
| No templates            | Shopify’s generated default | Shopify’s generated default | Shopify’s generated default |
| Only agents.md.liquid   | your agents.md.liquid       | your agents.md.liquid       | your agents.md.liquid       |
| Only llms.txt.liquid    | Shopify’s generated default | your llms.txt.liquid        | Shopify’s generated default |
| All three               | your agents.md.liquid       | your llms.txt.liquid        | your llms-full.txt.liquid   |

In Shopify’s words: “If no template is present for a given path, it falls back to your agents.md template, then to the Shopify-generated default.” Source: [Shopify.dev changelog: customize llms.txt, llms-full.txt and agents.md (28 May 2026)](https://shopify.dev/changelog/posts/customize-llmstxt-llms-fulltxt-and-agentsmd).

If you only want to change llms.txt, add `llms.txt.liquid` and leave the rest to Shopify. [Our Shopify guide](https://storewalk.ai/shopify) has the other settings.

On WooCommerce, the usual route is a plugin. Yoast SEO has a setting for it, "Under AI tools, go to LLMS.txt", and notes that "llms.txt is not available for multisite." The [WooCommerce guide](https://storewalk.ai/woocommerce) has the rest.

On BigCommerce we found no built-in setting for a store's llms.txt. BigCommerce publishes one for its own developer docs, which, as with the AI labs, is not the same thing. Third-party tools offer to generate one. See the [BigCommerce guide](https://storewalk.ai/bigcommerce).

On Magento and Adobe Commerce, several Marketplace extensions generate the file, and we found no setting in core. The [Magento guide](https://storewalk.ai/magento) covers the rest of the platform.

On any other site, it is a text file at the root. Serve it with a 200 status as `text/plain` or `text/markdown`, which is what our check's fix suggests, and make sure your server does not answer the path with an HTML page.

## What mistakes should you avoid?

Most of them come from treating the file as a marketing page.

- Writing it once and forgetting it. A collection that no longer exists is a dead link, and our check follows links.
- Stating terms the policy pages do not. If llms.txt says free returns and the returns page says 4.90 EUR, an agent has two answers and no way to choose.
- Pasting the whole catalogue. The spec's file is a summary; a list of four thousand product links is a sitemap, and you already have one.
- Blocking it. If robots.txt asks AI bots to stay out of the whole site, a well-behaved one will not read llms.txt either. [robots.txt for AI agents](https://storewalk.ai/guides/robots-txt-for-ai-agents) covers which bots to let in.

Once it is live, run a scan. The [agent readiness checklist](https://storewalk.ai/guides/agent-readiness-checklist) lists llms.txt among its cheap, optional tasks, beside the ones that matter more.

## Frequently asked questions

### Does Google use llms.txt?

No, for Search. Google's AI optimization guide says Google Search ignores AI text files such as llms.txt, and that adding one "will neither harm nor help" a site's rankings. Chrome's experimental Lighthouse Agentic Browsing category does check that one is present.

### Do ChatGPT, Claude or Perplexity read my llms.txt?

We found no statement from OpenAI, Anthropic or Perplexity that their assistants or crawlers read a site's llms.txt. They publish llms.txt files for their own documentation, which is a different thing.

### Does my Shopify store already have an llms.txt?

Yes. Shopify says it manages an agents.md for every store, and that /llms.txt and /llms-full.txt point to the same content by default. You can replace any of them with a template in your theme.

### What is the difference between llms.txt and robots.txt?

robots.txt asks bots where they may go. llms.txt tells a language model what your site is and which pages matter. One is a set of rules, the other a short guide, and a bot that obeys the first may never reach the second.

### How long should a store's llms.txt be?

Long enough to name your store, say what you sell and where you ship, and link your main collections and your policies. Our check wants at least 20 non-empty lines; the example in the builder gets there with fourteen links.

The sources below were last checked on the date shown.

## Sources

- [llmstxt.org: The /llms.txt file, v2](https://llmstxt.org/)
- [Google Search Central: AI optimization guide](https://developers.google.com/search/docs/fundamentals/ai-optimization-guide)
- [Chrome for Developers: Agentic Browsing scoring](https://developer.chrome.com/docs/lighthouse/agentic-browsing/scoring)
- [Shopify.dev changelog: customize llms.txt, llms-full.txt and agents.md (28 May 2026)](https://shopify.dev/changelog/posts/customize-llmstxt-llms-fulltxt-and-agentsmd)
- [Shopify.dev: agents.md.liquid](https://shopify.dev/docs/storefronts/themes/architecture/templates/agents-md-liquid)
- [Yoast: Enable llms.txt](https://yoast.com/help/enable-llmstxt/)
- [BigCommerce developer docs: AI agent setup](https://docs.bigcommerce.com/developer/docs/ai-agent-setup)
- [Your Store Wizards: llms.txt for BigCommerce stores](https://blog.yourstorewizards.com/llms-txt-generator-for-bigcommerce-stores/)
- [Adobe Commerce Marketplace: an llms.txt extension](https://commercemarketplace.adobe.com/mb-module-llms.html)
- [OpenAI: Overview of OpenAI crawlers](https://developers.openai.com/api/docs/bots)
- [Perplexity: Perplexity crawlers](https://docs.perplexity.ai/docs/resources/perplexity-crawlers)

Last checked 3 Nov 2026