llms.txt Best Practices for Ecommerce Stores and Marketplaces
For ecommerce and marketplaces, the best llms.txt is short: a clear store summary and links to category hubs, buying guides, shipping, returns and seller policies, never thousands of product URLs. Keep it accurate and in sync with your sitemap. Be realistic: llms.txt does not control access, and no major AI engine has confirmed using it for rankings. Product visibility in AI answers depends on crawlable pages, structured data that matches visible prices, reviews and consistent catalogue data.
Where llms.txt fits on a catalogue site
llms.txt is a proposed convention (llmstxt.org, September 2024) for a Markdown file at /llms.txt that summarises a site and links its key pages for language models. On a content site that might be twenty articles. On an ecommerce store or marketplace with tens of thousands of product pages, the real question is what to leave out.
Two facts shape the answer. The file does not control what crawlers can access; robots.txt and CDN or firewall settings do. And no major AI engine has confirmed using llms.txt for ranking or product recommendations; Google has said Search does not use it. So the file is a curated map for the tools and agents that choose to read it, not a way to push products into answers.
That does not make it pointless. Shopping agents and AI tools that do read it get a faster route to your policies and categories, and writing it is a useful check on whether your site structure makes sense to anyone outside the team.
What to include, and what to leave out
- A two-sentence summary: what you sell, to whom, in which countries, and whether you are a retailer or a marketplace of third-party sellers.
- Links to category hubs and top buying guides, not individual SKUs.
- Shipping, returns, warranty and payment pages, which answer the questions shoppers ask most before ordering.
- For marketplaces, seller rules, buyer protection and dispute policies.
- Size guides, compatibility charts and other reference pages that settle common questions.
Leave out product URLs that churn, faceted and filtered pages, internal search results, cart and account pages, and anything blocked in robots.txt. A file that lists thousands of URLs defeats its own purpose; your sitemap already handles discovery.
| Section | Include | Leave out |
|---|---|---|
| Summary | What you sell, where, retailer or marketplace | Slogans and promotions |
| Catalogue | Category hubs and top buying guides | Individual SKUs, filters, search results |
| Policies | Shipping, returns, warranty, payment | Outdated policy summaries |
| Marketplace trust | Seller rules, buyer protection, disputes | Internal seller tools |
| Reference | Size guides and compatibility charts | Cart, account and checkout pages |
Keeping it accurate as the catalogue changes
Catalogue sites change daily, so an llms.txt goes stale quickly unless someone owns it. Generate it from the same source as your navigation or sitemap, so a renamed category or retired guide updates automatically. Check on a schedule that every linked URL returns a 200 status, is indexable and is not redirected.
Watch policy pages especially. If your returns window changes and the linked page is updated but the summary in the file is not, you have created exactly the inconsistency that leads AI tools to quote the wrong terms to shoppers, and the wrong terms then turn into support tickets and refund disputes.
For multi-country stores, decide whether one file covers all markets or each country site has its own. If prices, shipping and returns differ by market, separate files at each country root are clearer and reduce the risk of one market's terms being quoted in another.
Signals that matter more for product answers
When a shopper asks an assistant which product to buy, the answer draws on pages the engine can fetch and trust. For a store, that means product and category pages crawlable by Googlebot and by AI citation crawlers such as OAI-SearchBot and PerplexityBot; structured data for products, prices, availability and reviews that matches what is visible on the page; and consistent product names and identifiers across your site, feeds and marketplace listings.
Third-party coverage matters too: buying guides, review publications and comparison sites are cited heavily for best-product questions. No llms.txt can substitute for any of this.
Product feeds also matter wherever engines use them for shopping results. Keep feed titles, prices and availability identical to the product pages, because mismatches can lead to products being shown at the wrong price or dropped altogether.
| Signal | Why AI engines use it | Typical owner |
|---|---|---|
| Crawlable product and category pages | Engines can only cite what they can fetch | Engineering |
| Structured data matching visible content | Clear prices, availability and reviews | Ecommerce and SEO |
| Consistent names and identifiers | Avoids merged or confused products | Catalogue team |
| Reviews and buying-guide coverage | Often cited for best-product questions | PR and partnerships |
| Clear policy pages | Answers pre-purchase questions | Customer service |
Worked example: a marketplace for trading tools and courses
Picture a marketplace selling third-party trading courses, indicators and platform add-ons to retail traders. Its llms.txt would link category hubs for courses, indicators and platform tools, a buyer protection page, seller vetting rules, the refund policy and a guide to choosing a course. It would not list each of its thousands of listings.
For its products to appear when traders ask AI where to learn forex or which indicator suits a strategy, the real work is elsewhere: indexable listing pages with clear descriptions and honest reviews, consistent seller and product names, and coverage on publications that review trading tools. Because trading products sit close to regulated financial promotion, listing claims must be accurate and free of guaranteed-return language.
What Suggesting.ai does for ecommerce and marketplaces
Suggesting.ai's free audit, delivered in 48 hours, tests the product and category questions your shoppers ask across ChatGPT, Perplexity, Google AI features and Copilot. It records which stores and products are recommended, which sources are cited, whether AI crawlers can reach your catalogue past CDN and firewall rules, and whether your product data is consistent. It also reviews your llms.txt for accuracy.
We then run organic GEO on catalogue structure, structured data, entity consistency and citations, and manage ChatGPT Ads where available, since sponsored placements matched to shopping conversations can reach buyers while organic visibility builds. When an assistant is suggesting where to buy, the aim is that your store is on the list.
Measuring what matters
Check server logs for /llms.txt requests by user agent; that shows whether the file is read at all. Measure commercial impact through AI referral sessions and orders, where ChatGPT links carry utm_source=chatgpt.com, product mention rate across a fixed prompt set, and the accuracy of prices and policies stated in answers.
Over time, compare product mention rates in your top categories against your main competitors, and check that AI answers quote your current prices and policies.
Frequently asked questions
Should I list every product in llms.txt?
No. The file is meant to be a curated summary. On large catalogues, link category hubs, buying guides and policy pages instead. Product URLs change too often, and a file with thousands of links is no more useful than your sitemap, which already handles discovery for search engines.
Will llms.txt get my products into ChatGPT shopping answers?
There is no evidence that it will. No major AI engine has confirmed using llms.txt for recommendations. Product visibility depends on crawlable pages, accurate structured data, consistent product data, reviews and coverage on the sites engines cite. Treat llms.txt as low-cost hygiene alongside that work.
How often should an ecommerce llms.txt be updated?
Whenever a linked category, guide or policy changes. The easiest approach is to generate it automatically from the same source as your navigation or sitemap, then check the linked URLs for errors and redirects on a schedule, for example weekly, so the file never drifts from the live site.
Do marketplaces need different llms.txt content from retailers?
Yes. Marketplaces should explain that products come from third-party sellers and link seller rules, buyer protection and dispute pages, because shoppers and AI tools ask about trust and refunds. Retailers focus more on shipping, returns and warranty, since they sell and fulfil the products themselves.
Can llms.txt stop AI from scraping my product data?
No. llms.txt is informational. Use robots.txt for crawlers that respect it and CDN or firewall bot rules for enforcement. Blocking citation crawlers also removes you from AI answers, so most stores allow them and make a separate decision about training crawlers such as GPTBot.
Want AI to suggest your brand instead of a competitor?
Want to know which stores AI recommends for your top products? Request Suggesting.ai's free 48-hour audit of your catalogue's AI visibility.