In recent years, the rapid evolution of generative AI and large language models (LLMs) has begun to reshape how the internet is used. Traditionally, people opened a browser, typed keywords into a search engine, and found information themselves. Today, however, "AI agents" are emerging — systems that collect information on behalf of users and support their decisions and tasks.

In response to this shift, Shopify has introduced a new mechanism for making store information more accessible to AI agents. Shopify stores now automatically generate three AI agent discovery URLs: /agents.md, /llms.txt, and /llms-full.txt.

This article takes a practical look at the concept behind llms.txt, what Shopify's native support means, and how e-commerce sites should approach information design going forward.

What is llms.txt? Information delivery for the AI agent era

llms.txt is a plain-text site overview file designed to help LLMs and AI agents efficiently understand a website's content. It is a growing initiative aimed at making information accessible in the age of AI agents.

According to llmstxt.org, the file that proposed the standard, it serves as a signpost that helps AI quickly identify which pages on a site contain the most important content.

Think of it like an introductory letter you hand to a new business contact, or an official project briefing document — a concise orientation for whoever (or whatever) is arriving for the first time.

It is not yet a fully established standard like robots.txt or sitemap.xml. It remains a "proposed standard," but it is attracting growing interest among AI companies and platforms.

The table below compares existing web standards with the role of llms.txt:

Where traditional files define the scope of crawling or a list of URLs, llms.txt is distinct in that it aims to communicate the meaning and context of information to AI.

Shopify's automatically served agent discovery files

Shopify has been quick to act on this trend, rolling out an update that automatically serves the following three URLs for every Shopify store — collectively described as "agent discovery files":

• /agents.md
• /llms.txt
• /llms-full.txt

Accessing these URLs returns the store's name and basic information, main product categories, links to policies, and more — all in plain Markdown or text format that AI can readily parse.

If you run a Shopify store, try visiting your store URL with /llms.txt appended — for example, https://your-store.com/llms.txt — to see exactly what AI agent-facing information your store is currently serving.

When we accessed our own site's /llms.txt, we found an AI agent-facing description in plain text, as shown in the image below. It includes not only basic store information, but also instructions for how AI agents can interact with the store, which URLs to reference, how to retrieve product information, and notes on checkout behavior.

Among these files, agents.md serves as the primary description document that AI agents use to understand the store. Per Shopify's specification, merchants can fully customize the content of /agents.md, /llms.txt, and /llms-full.txt by adding dedicated custom templates to the store theme. If no custom template is provided, Shopify automatically generates default content from the store's basic settings and applies it as agents.md.

Full implementation details are available in the official Shopify developer documentation (Shopify.dev).

Shopify's track record of bringing emerging technology to production

For those who have been watching Shopify's development philosophy, its early adoption of this proposed standard will come as no surprise. Shopify has consistently taken new web technologies and elevated them into production-ready platform features.

Examples include its early and central adoption of GraphQL as the foundation of its API, as well as its incorporation of headless commerce, server-side rendering, edge delivery, and modern React-based frontend development — all made accessible within a real commerce operations context.

This latest support for llms.txt and agents.md is another example of Shopify's forward-looking approach, anticipating the growing role of AI agents in the commerce landscape.

Flagship's perspective and approach

At Flagship, we had actually begun prototyping a dedicated app that would make it easy for Shopify merchants to generate and manage llms.txt files.

However, our product team, drawing on Shopify's track record of rapid adoption of technology trends, concluded that "there's a real possibility Shopify will eventually support this natively as a standard feature." As a result, we chose not to release the app publicly.

Shopify's subsequent move aligned closely with that assessment. The ability to distinguish between what a platform will eventually provide natively and what should be addressed through customization is a critical perspective in long-term system design.

The lens you need for designing e-commerce sites in the AI agent era

Web design and e-commerce operations going forward require us to confront a new reality: the audience for a site's information is expanding from two parties — human users and search engines — to three, with AI agents joining as a distinct third category.

It's increasingly plausible that instead of visiting multiple e-commerce sites to compare products themselves, users will ask an AI assistant: "Within a budget of ¥20,000, find and suggest shoes that would go with what I already own." In that scenario, AI agents would read each store's llms.txt, agents.md, and product data to surface the best recommendations for the user.

What is becoming increasingly important in this context is how well you can deliver accurate, organized, and structured official information to AI agents.

That said, there are important caveats to keep in mind in practice:

• Simply placing llms.txt or agents.md files does not complete all AI optimization (AIEO) work.
• It is essential to comprehensively address product information throughout the site, structured data (e.g., JSON-LD), meta information, and logical site architecture.
• At this point, not all AI agents make use of files like llms.txt in the same way.

The goal should not be to add new files for their own sake, but to treat them as one part of a broader, systematic approach to information architecture across the entire site.

Closing: preparing for the future of the buying experience

Shopify's support for llms.txt and agents.md represents a first step toward the AI agent-powered commerce of the future. E-commerce operators will increasingly need to pursue not only visually polished experiences for human visitors, but also information transparency and structure that AI can readily interpret.

As professionals specializing in advanced Shopify customization and system development, Flagship keeps a close eye on these technology trends and their practical value. We go beyond building screens — we optimize data structures and information architecture with the AI era in mind, supporting commerce strategies that stay one step ahead.