If you run a WooCommerce store, someone has probably told you to add an llms.txt file "before AI search leaves you behind." So what is llms.txt, really? It is a plain Markdown file that sits at the root of your site and gives large language models a short, curated map of your most useful pages. It takes about 30 minutes to write by hand for a typical store with a few hundred products. The stakes are low in both directions. It costs almost nothing to publish, and it will not rescue a store that has thin product pages. This guide explains where the idea came from, the exact format, where the file goes, how it compares with robots.txt and sitemap.xml, and a complete example for an online store. It also gives you an honest answer on SEO, based on what Google has actually published rather than what tool vendors hope is true.
Key takeaways
- llms.txt is a proposed standard, published by Jeremy Howard of Answer.AI in September 2024, for a Markdown file at
/llms.txtthat summarises a website for language models. - The format is simple: one H1 with the site name, a short blockquote summary, optional notes, then H2 sections that list Markdown links. A section called "Optional" holds pages that can be skipped.
- It is not a ranking factor. Google says you do not need special AI text files to appear in its AI features.
- For an online store, the useful content is policies, sizing, shipping, returns and your main category pages, not every product URL.
- Treat it as cheap housekeeping. Spend most of your effort on clear product pages, which help shoppers, search engines and AI tools at the same time.
What is llms.txt, in plain terms?
llms.txt is a text file written in Markdown, placed at the root of a website, that tells large language models what the site is about and where its most important content lives. It is a proposal, not an official web standard, and no search engine or AI company is required to read it.
A large language model (LLM) is the kind of system behind ChatGPT, Claude, Gemini, Perplexity and Microsoft Copilot. When one of these tools browses the web to answer a question, it has to deal with real web pages. Those pages are full of menus, cookie banners, scripts, footers and ads. A model also has a limited context window, which is the amount of text it can consider at once. A long, messy HTML page eats that budget quickly.
The llms.txt idea tackles this with a short, hand-picked summary. Instead of crawling your whole store, a tool could read one small file and learn three things fast: who you are, what you sell, and which pages hold the facts worth quoting. Think of it as the note you would leave for a smart new employee on their first day, written in a format machines parse easily.
Who created llms.txt, and is it an official standard?
Jeremy Howard, co-founder of Answer.AI and fast.ai, proposed llms.txt in September 2024 and published the specification at llmstxt.org. It is an open community proposal. It has not been adopted by a standards body such as the IETF or W3C, and it has no formal approval from Google, OpenAI or Anthropic.
The full proposal lives at llmstxt.org. Its original focus was documentation. Developers using AI coding assistants wanted a fast way to point a model at the right parts of a software library, without the model wading through hundreds of HTML pages. Many developer documentation sites adopted the file early for exactly that reason.
What does the llms.txt format look like?
The llms.txt format uses ordinary Markdown in a fixed order: an H1 with the site or project name (the only required part), a blockquote with a short summary, optional paragraphs or lists of extra detail, then zero or more H2 sections, each containing a list of Markdown links with optional notes.
Here is the structure the specification describes, in order:
- H1 title. A single line starting with
#and your site name. This is the only mandatory element. - Blockquote summary. A line starting with
>that sums up the site in a sentence or two. It should contain the key facts a model needs to understand everything below it. - Extra detail. Optional plain paragraphs or bullet lists with no headings. Use these for context such as who you serve or important caveats.
- H2 file lists. Sections starting with
##. Each holds a bulleted list where every item is a Markdown link, written as[Page name](URL), optionally followed by a colon and a short note. - An "Optional" section. An H2 named exactly "Optional" has a special meaning. Links there are secondary and can be skipped when a tool needs a shorter context.
A minimal valid file looks like this:
# Example Coffee Co
> Small-batch coffee roaster selling whole beans and ground coffee online, shipping within the United States.
## Policies
- [Shipping](https://example.com/shipping/): Rates, carriers and delivery times
- [Returns](https://example.com/returns/): 30-day returns on unopened bags
Note that the file is Markdown, not a strict data format like JSON or XML. That is deliberate. Language models read Markdown well, and humans can write it in any text editor. The specification also suggests a related idea: offering a clean Markdown copy of important pages at the same address with .md added to the end. Few WordPress stores do this today, and you do not need it to publish a useful llms.txt.
What is llms-full.txt, and do you need it too?
llms-full.txt is a companion file that some sites publish alongside llms.txt. Instead of links, it contains the full text of the important pages joined into one large Markdown document. It grew up as a common convention around the proposal, mostly on documentation sites, rather than as a core required part of the main format.
The difference is easy to picture. llms.txt is a table of contents. llms-full.txt is the whole book. A tool that reads llms.txt still has to fetch each linked page. A tool that reads llms-full.txt gets everything in one request.
For most online stores, llms-full.txt is not worth the upkeep. Here is a worked example. Say your store has 400 products, and each product page has 250 words of description. That is 100,000 words of product copy alone. Add policies and guides and you are well past what many tools would read in one pass. Every price change or stock update would also make the file stale, unless something regenerates it automatically.
Where llms-full.txt can make sense is a small, stable body of text. A store with ten pages of care guides, sizing advice and policies could join those into one file of a few thousand words. In my opinion, that is the upper limit of what is worth maintaining by hand. Beyond that, keep to llms.txt and let it link out.
Where do you place the llms.txt file on your website?
Place llms.txt in the root directory of your domain, so it loads at https://yourstore.com/llms.txt. The specification uses the root path, the same way robots.txt works. You do not submit it anywhere. There is no "llms.txt submission" in Google Search Console or any AI tool.
On a WordPress site, the root is usually the folder that holds wp-config.php, often called public_html or www on shared hosting. A file uploaded there loads at the top level of your domain.
Two details trip people up. The name is plural, llms.txt, even though many people search for "llm.txt". And subdomains are separate, so shop.example.com needs its own file. If an update does not show, clear your page cache and CDN cache.
To check any site's file, including a competitor's, type their domain followed by /llms.txt into your browser. If you see Markdown text, they have one. If you see a 404 page, they do not.
How is llms.txt different from robots.txt and sitemap.xml?
robots.txt tells crawlers what they may and may not fetch. sitemap.xml lists every URL you want search engines to discover. llms.txt is a short, human-curated summary that suggests which pages matter most for a language model. None of them replaces the others.
| Feature | robots.txt | sitemap.xml | llms.txt |
|---|---|---|---|
| Main job | Allow or block crawling | List URLs for discovery | Summarise and point to key content |
| Format | Plain text directives | XML | Markdown |
| Written for | Crawlers | Search engines | Language models and people |
| Status | IETF standard (RFC 9309) | Widely supported protocol | Community proposal |
| Covers | Rules for paths | Ideally every indexable URL | A curated handful of pages |
| Controls access | Yes, for crawlers that respect it | No | No |
| Used by Google Search | Yes | Yes | No, per Google's guidance |
| Location | /robots.txt | Any path, often /sitemap.xml | /llms.txt |
The most important row is "Controls access." If you want to stop an AI crawler from using your content, llms.txt will not do it. You would add rules to robots.txt for the user agents those companies publish, such as GPTBot for OpenAI. Even then, robots.txt is a request that well-behaved crawlers honour, not a lock.
Will llms.txt help your SEO?
Not in Google Search, based on Google's own guidance. Google states that you do not need to create new machine-readable files or AI text files to appear in AI Overviews or AI Mode, and the normal SEO basics apply. Any benefit from llms.txt today is speculative and comes from other AI tools, not Google rankings.
Google's documentation page AI features and your website covers how sites appear in AI Overviews and AI Mode. It says the same SEO best practices apply as for Search generally, and that no special files or markup are required. That rules out llms.txt as a Google ranking factor or a requirement for Google's AI answers.
What about other AI tools? Here is the honest position. Some AI companies publish llms.txt files for their own documentation. That is not the same as their crawlers reading other people's llms.txt files when answering questions. At the time of writing, I could not verify a public, official statement from OpenAI, Anthropic, Perplexity or Microsoft saying their answer engines use llms.txt as an input. Claims that "ChatGPT reads your llms.txt" should be treated as unproven unless they link to such a statement.
There are narrower real uses. A person can paste your llms.txt URL into an AI assistant and ask it to read the linked pages, and some developer and agent tools fetch these files. Writing one also forces a useful review of your key pages.
My opinion, labelled as such: publish an llms.txt because it is cheap and harmless, not because it will move rankings. If someone sells you an llms.txt service with traffic promises, ask for evidence you can check. The work that reliably helps in both classic search and AI answers is still the old work. That means fast pages, clear product titles, real descriptions, accurate structured data and policy pages that answer the question in the first paragraph.
What should an llms.txt for an online store include?
An llms.txt for an online store should cover who you are and what you sell, where you ship, the policies shoppers ask about most (shipping, returns, warranty, payment), sizing or compatibility guides, your main category pages and a contact page. Individual products only belong there if you sell a small catalogue.
What does a complete llms.txt example for a WooCommerce store look like?
A complete llms.txt for a WooCommerce store usually runs 20 to 40 lines: the H1 store name, a one-sentence blockquote, a short paragraph of key facts, then H2 sections for policies, guides, categories and contact, with bestsellers under Optional. The example below is for a fictional store.
The store below is invented for illustration. The domain uses example.com, which is reserved for documentation.
# Northfield Outdoor Supply
> Northfield Outdoor Supply is an online store selling hiking boots, rain jackets and camping gear for adults, shipping across the United States and Canada.
- Prices are in US dollars. Canadian orders are charged in US dollars.
- Orders ship from one warehouse in Vermont, usually within 2 business days.
- All products are new. We do not sell used or refurbished gear.
## Policies
- [Shipping](https://example.com/shipping/): Carriers, delivery times and rates for the US and Canada
- [Returns and exchanges](https://example.com/returns/): 30-day returns on unworn items, how to start a return
- [Warranty](https://example.com/warranty/): What the one-year warranty covers and how to claim
- [Payment methods](https://example.com/payment/): Accepted cards and wallets
## Buying guides
- [Boot size guide](https://example.com/boot-size-guide/): US, UK and EU sizes with foot length in cm and inches
- [Choosing a rain jacket](https://example.com/rain-jacket-guide/): Waterproof ratings explained in plain language
- [Care instructions](https://example.com/care/): Cleaning and waterproofing boots and jackets
## Shop by category
- [Hiking boots](https://example.com/product-category/hiking-boots/): Men's and women's boots for day hikes and backpacking
- [Rain jackets](https://example.com/product-category/rain-jackets/): Lightweight and insulated waterproof jackets
- [Camping gear](https://example.com/product-category/camping/): Tents, sleeping bags and stoves
## About and contact
- [About us](https://example.com/about/): Who runs the store and where it is based
- [Contact](https://example.com/contact/): Email address, support hours and response times
- [FAQ](https://example.com/faq/): Answers to common order questions
## Optional
- [Bestselling boots](https://example.com/product/trail-mid-boot/): Our most popular mid-height hiking boot
- [Blog](https://example.com/blog/): Trail guides and gear reviews
How do you create an llms.txt file, step by step?
To create llms.txt, list the 10 to 25 pages that answer most customer questions, write an H1 and one-sentence summary, group the links under H2 headings with short notes, save it as UTF-8 plain text named llms.txt, upload it to your site root and open it in a browser to confirm it loads.
Option 1: Write it by hand
- Collect your pages. Open your WordPress menu, footer and policy pages. Write down the URLs that cover shipping, returns, sizing, categories and contact.
- Write the summary. One sentence that says what you sell, to whom and where you ship. Avoid slogans. "Quality you can trust" tells a model nothing.
- Add key facts. Two to five short bullets with currency, shipping origin, and anything unusual, such as "made to order, ships in 3 weeks."
- Group the links. Use H2 headings like Policies, Buying guides, Shop by category and About and contact. Add a one-line note to each link.
- Add an Optional section. Put nice-to-have pages there, such as the blog or bestsellers.
- Save correctly. Use a plain text editor such as Notepad, TextEdit in plain text mode, or VS Code. Name the file
llms.txtexactly. - Upload to the root. Use your host's file manager or an SFTP client such as FileZilla, and place the file next to
wp-config.php. - Test it. Visit
https://yourstore.com/llms.txt. Then click a few of the links to confirm none return a 404.
Option 2: Use a free generator
If you would rather not format Markdown yourself, the free llms.txt generator on this site formats the file for you. Then upload it as in steps 7 and 8 above. Always read the result before publishing. A generator saves typing, but you know which pages matter.
Option 3: Generate it inside WordPress
Some WordPress plugins write the file for you. StorePinger Pro, for example, includes an AI visibility feature that tracks visits from AI tools such as ChatGPT, Perplexity, Gemini, Claude and Copilot, and writes an llms.txt file for the store. Pairing the two is useful, because you can publish the file and then watch whether AI referrals change, instead of guessing. You can see what is included on the StorePinger features page and the pricing page. A plugin is not required, though. A hand-written file works the same way.
What are the common llms.txt mistakes to avoid?
The most common llms.txt mistakes are naming it llm.txt, placing it in a subfolder, dumping the whole sitemap into it, linking to cart or account pages, using relative URLs, letting links go stale after a redesign, and believing it blocks AI crawlers. Each one is easy to fix in minutes.
- Wrong file name. The proposal uses
llms.txtwith an s. Check the spelling in the URL bar. - Wrong location. A file at
/wp-content/uploads/llms.txtis not at the root. Media Library uploads land there, so do not upload it through WordPress media. - Too many links. Hundreds of product links turn a summary into a second sitemap. Curate.
- Stale content. If you change your returns window from 30 to 14 days, update the note in llms.txt the same day. Wrong facts are worse than none.
- Treating it as a block list. To restrict crawlers, use robots.txt. llms.txt has no blocking power.
- Expecting rankings. As covered above, Google does not use it for Search. Set expectations accordingly.
What are the llms.txt best practices for stores?
The best llms.txt files for stores are short, factual and current. Keep it under about 50 links, lead with a plain one-sentence summary, put policies first, describe every link in a few words, review it whenever policies or categories change, and make sure the linked pages themselves answer questions clearly.
That last point matters most. llms.txt only points at pages. If your returns page opens with three paragraphs of brand story before mentioning the return window, the file has pointed a model at a weak answer. Rewrite the page so the first sentence answers the question, for example "You can return unworn items within 30 days for a full refund."
A few more practices worth adopting:
- Match your structured data. If your product schema says one shipping price and your llms.txt says another, you are sending mixed signals. Keep facts consistent across pages, schema and the file.
- Watch real data. Track referrals from AI tools over time. If the numbers stay tiny for your niche, do not spend more time on the file.
If you are working on store growth more broadly, the bigger wins usually sit elsewhere. Recovering shoppers who leave at checkout is a good example, and our guide on how to reduce cart abandonment covers the fixes that tend to matter most. You can also read about abandoned cart recovery for WooCommerce and see abandoned cart email examples.
Agencies managing many client stores can reuse one template and a quarterly link check. The StorePinger agencies page explains multi-site setups.
Conclusion: what is llms.txt worth to your store?
So, what is llms.txt in the end? It is a short Markdown file at your site root that offers language models a curated map of your best content. It came from Jeremy Howard and Answer.AI in September 2024. It has a simple format, a companion convention called llms-full.txt, and no role in Google rankings, according to Google's own guidance. Whether other AI tools read it widely is still an open question.
The sensible move is to treat it as low-cost housekeeping. Write a clear file of 20 to 40 lines, point it at your policies, guides and categories, upload it to the root and check it each quarter. Then put your real energy into product and policy pages that answer questions in the first sentence.
Your one next step: open the free llms.txt generator, build your file in a few minutes and publish it today. If you also want to see whether AI tools send visitors, StorePinger can track traffic sources inside WordPress, including AI referrals, and you can try it on the live demo.