If you have been following discussions around AI search, GEO, AEO, or making websites “AI-friendly,” you may have started seeing two unfamiliar files mentioned repeatedly:
llms.txt
and
llms-full.txt
Some articles describe them as the AI equivalent of robots.txt. Others suggest that adding these files can improve your chances of appearing in ChatGPT, Google AI Overviews, Claude, or other AI-generated answers.
That description is too simplistic.
llms.txt is a proposed way to give AI agents a cleaner, more useful path through your website. It has legitimate use cases, particularly for documentation-heavy websites.
But it is not a replacement for robots.txt, not a replacement for your XML sitemap, and not an AI search ranking shortcut.
Google has explicitly stated that it does not use llms.txt for Google Search, including its generative AI features. Creating one neither improves nor damages your Google rankings.
So what exactly does llms.txt do? What is llms-full.txt? And should you add either of them to your website?
Here is what website owners need to know.
What Is llms.txt?
llms.txt is a Markdown-formatted file designed to help AI agents understand and navigate a website.
The proposal was originally published by Jeremy Howard in September 2024 and was updated to version 2 in August 2026. The current proposal describes llms.txt as a conventional location where a website can provide background information, instructions, and links to useful resources in a format that is easy for language models and agents to process.
A typical file is available at:
https://example.com/llms.txt
The important idea is not that the file contains your entire website.
Instead, it acts more like a curated guide.
Suppose you run a WordPress plugin company. Your website may contain:
- Product pages
- Pricing pages
- Documentation
- Tutorials
- Blog posts
- Changelogs
- Account pages
- Affiliate pages
- Legal pages
- Hundreds of other URLs
An AI agent trying to understand your product does not necessarily need all of those pages.
Your llms.txt file can point it toward the pages that provide the most useful context.
For example:
# ExamplePlugin
> ExamplePlugin is a WordPress plugin for building product comparison tables and recommendation sections.
## Product
- [Features](https://example.com/features/): Overview of the main features
- [Pricing](https://example.com/pricing/): Available plans and pricing
## Documentation
- [Getting Started](https://example.com/docs/getting-started/): Installation and initial setup
- [Product Tables](https://example.com/docs/product-tables/): How to build product tables
- [Filters](https://example.com/docs/filters/): How product filtering works
## Optional
- [Blog](https://example.com/blog/): Tutorials and industry articles
An agent that understands the convention can read the file, identify the relevant section, and retrieve the pages it needs instead of trying to explore the whole website.
Why Was llms.txt Proposed?
Modern web pages are built primarily for humans and browsers.
A page may contain the actual article or documentation alongside:
- Navigation menus
- Cookie banners
- Popups
- Related-post sections
- Advertisements
- JavaScript
- CSS
- Tracking scripts
- Headers
- Footers
- Sidebars
A browser can turn all of this into a usable page for a human.
An AI agent often wants something much simpler: the actual information.
The llms.txt proposal tries to solve two related problems.
1. Finding the Right Information
A large website may contain thousands of URLs.
An AI agent should not have to inspect every page just to find the documentation explaining one feature.
llms.txt gives the agent a curated index.
2. Getting Cleaner Content
The current specification also recommends providing Markdown versions of useful pages where practical.
For example:
example.com/docs/setup/
could have a machine-friendly version such as:
example.com/docs/setup.md
Markdown removes much of the presentation layer and leaves headings, text, lists, links, tables, and code in a comparatively clean format.
The v2 proposal specifically recommends mechanisms that can point agents toward Markdown alternatives and toward the appropriate llms.txt file for a section of a website.
What Should an llms.txt File Contain?
The specification is intentionally simple.
An llms.txt file can contain:
H1: Website or Project Name
For example:
# ExamplePlugin
The H1 is the only required part of the current specification.
A Short Summary
A blockquote can explain what the company, website, product, or project actually is.
> ExamplePlugin helps WooCommerce stores create product bundles and personalized offers.
This gives the agent immediate context.
Additional Instructions or Context
You can provide useful information that helps an agent interpret your site correctly.
For example, a software company might explain which version of its documentation is current.
H2 Sections Containing Useful Links
You can organize URLs into sections such as:
## Product
## Documentation
## API Reference
## Tutorials
## Support
Each item is normally a Markdown link with an optional short explanation.
The proposal also uses an Optional section convention for secondary resources that an agent can ignore when it needs to keep its context small.
What Is llms-full.txt?
This is where the terminology becomes confusing.
llms-full.txt generally contains the actual contents of many or all relevant pages in one large file.
Think of the difference this way:
llms.txt tells an agent where the useful information is.
llms-full.txt attempts to give the agent the useful information directly in one file.
For example, your llms.txt might contain:
## Documentation
- Getting Started
- Product Settings
- Integrations
- Troubleshooting
An llms-full.txt file could contain the full text of all four documentation sections.
Platforms such as Mintlify automatically generate both types of files. Mintlify describes llms-full.txt as a single file containing an entire documentation site so AI tools can load it as context. Cloudflare uses the same approach and says its llms-full.txt files are intended for uses such as offline indexing, bulk vectorization, and large-context models.
Is llms-full.txt Part of the Official llms.txt Specification?
This distinction matters.
The current llmstxt.org v2 specification defines llms.txt, but it does not define llms-full.txt as a required companion file. llms-full.txt has become an ecosystem convention through documentation platforms and real-world implementations rather than a requirement of the core proposal.
So you should not assume that implementing llms.txt means you must also create llms-full.txt.
llms.txt vs llms-full.txt
The difference is easier to understand in a table.
| File | Main Purpose | Usually Contains |
|---|---|---|
llms.txt | Help an agent discover useful information | Summaries, sections, links and descriptions |
llms-full.txt | Give an agent large amounts of content directly | Full text from many documentation pages |
A useful analogy is:
llms.txt = table of contents
llms-full.txt = the book
That does not mean the second option is automatically better.
For a website with thousands of documentation pages, putting everything into one file can create a very large resource. An agent may only need three pages to answer a question.
In that situation, a smaller index that lets the agent retrieve only the necessary pages can be more practical.
Is llms.txt the Same as robots.txt?
No.
They serve very different purposes.
robots.txt Is About Crawler Access
robots.txt communicates crawling instructions to automated crawlers.
For example, you might use it to tell certain crawlers not to access a directory.
llms.txt Is About Context and Discovery
llms.txt does not block an AI crawler.
It does not grant special permission.
It does not prevent your content from being used.
It simply provides information and links that compatible agents can use.
The llms.txt specification itself explains that the proposal is intended to coexist with mechanisms such as robots.txt and XML sitemaps rather than replace them.
If your concern is controlling whether particular crawlers can access your site, llms.txt is not the appropriate mechanism.
Is llms.txt the Same as sitemap.xml?
No.
A sitemap is primarily about URL discovery.
A website might have an XML sitemap containing thousands of pages:
/product-a/
/product-b/
/blog/article-1/
/blog/article-2/
/category/news/
/author/john/
/documentation/example/
...
That is useful for search engines.
But an AI agent trying to understand a specific product may not need every indexable URL on your site.
llms.txt is supposed to be more selective.
Instead of saying:
Here are 6,000 URLs.
you can effectively say:
Here are the 20 resources that explain this website and product most clearly.
The official proposal specifically positions llms.txt as complementary to sitemaps rather than a replacement for them.
Does llms.txt Improve SEO?
There is currently no basis for claiming that llms.txt improves Google rankings.
Google clarified this directly in June 2026.
Its documentation states that you do not need AI-specific machine-readable files such as llms.txt to appear in Google Search or Google’s generative AI features.
Google Search does not use the file specially, and maintaining one will neither positively nor negatively affect your visibility or rankings in Google Search.
That means claims such as:
“Add llms.txt to rank in AI Overviews.”
or:
“You need llms.txt for Google AI Mode.”
are not supported by Google’s guidance.
Does llms.txt Help You Appear in ChatGPT, Claude, or Other AI Answers?
This needs a more careful answer.
Publishing an llms.txt file does not guarantee that an AI assistant will read it, use it, cite your website, or recommend your product.
There is no universal requirement that every AI crawler, answer engine, or agent must process llms.txt.
At the same time, the convention has gained practical adoption, particularly around software documentation and agent workflows. The current v2 proposal specifically discusses coding agents using these files to find API references and tutorials. Documentation platforms such as Mintlify generate them automatically, while companies such as Cloudflare maintain their own implementations.
The sensible expectation is therefore:
llms.txt can make your content easier to navigate for systems that choose to use it. It cannot force an AI system to use or cite your content.
That distinction prevents a useful technical convention from turning into another SEO myth.
Who Should Consider Creating llms.txt?
Not every website needs one.
Documentation Websites
This is probably the strongest use case.
If you maintain:
- API documentation
- Plugin documentation
- SDK documentation
- Developer guides
- Technical tutorials
- Product knowledge bases
giving agents a clean documentation map can be useful.
SaaS and Software Product Websites
Software products often have information distributed across features, documentation, integrations, pricing, API references, and tutorials.
A curated llms.txt can identify the authoritative resources.
WordPress Plugin and Theme Companies
A WordPress product company may have hundreds of tutorials and documentation pages.
Instead of listing everything, an llms.txt could organize resources such as:
## Products
## Features
## Documentation
## Integrations
## Tutorials
## Pricing
That provides considerably better context than dumping every URL from the sitemap.
Large Educational or Knowledge Websites
Sites with extensive structured educational material may also benefit when agents need a clear route into their content.
Who Probably Does Not Need llms.txt?
If you operate a five-page local business website containing:
- Home
- About
- Services
- Contact
- Blog
creating an llms.txt file is unlikely to be a meaningful priority.
The site is already small enough for agents and search engines to navigate.
You should spend your time first on things such as:
- Clear website architecture
- Useful content
- Internal linking
- Crawlability
- Accurate titles and descriptions
- Structured data where appropriate
- XML sitemaps
- Correct
robots.txtconfiguration - Fast, accessible pages
Google’s current guidance for generative AI search continues to emphasize the fundamentals: crawlable pages, solid technical SEO, useful original content, and a good page experience rather than special “AI SEO” files or hacks.
Do You Need llms-full.txt?
For most normal websites, probably not.
llms-full.txt makes the most sense when you have a substantial body of documentation that someone may genuinely want to ingest in bulk.
For example:
Good potential use case:
A developer platform with hundreds of API and SDK documentation pages.
Less useful use case:
A restaurant website with 12 pages.
Cloudflare’s implementation illustrates the intended type of workload well. It provides llms-full.txt for its documentation and describes use cases including offline indexing and bulk vectorization.
If your documentation platform automatically generates the file, there may be little downside to making it available.
Manually maintaining a giant duplicate copy of your entire website is a different matter.
How to Create an llms.txt File
If you decide it makes sense for your website, keep the implementation simple.
Step 1: Identify the Most Important Resources
Do not automatically copy your XML sitemap.
Choose resources that help someone understand your website.
For a software product, that might mean:
- Product overview
- Important features
- Pricing
- Getting started guide
- Documentation
- Integrations
- API documentation
- Troubleshooting
- Important policies
Step 2: Write a Clear Description
Explain what your website or product actually does.
Avoid marketing language such as:
The world’s most innovative and revolutionary solution.
Instead write:
ExamplePlugin is a WooCommerce plugin for managing inventory across multiple store locations.
That sentence gives an agent useful information.
Step 3: Organize Links Into Sections
Group links according to purpose.
For example:
## Product
- [Product Overview](https://example.com/product/): Main product information
- [Features](https://example.com/features/): Detailed feature list
- [Pricing](https://example.com/pricing/): Plans and licensing
## Documentation
- [Getting Started](https://example.com/docs/getting-started/): Installation and setup
- [Inventory Management](https://example.com/docs/inventory/): Inventory documentation
## Support
- [Knowledge Base](https://example.com/help/): Help articles
Step 4: Publish It at an Accessible URL
For a site-wide implementation, the usual URL is:
example.com/llms.txt
The v2 proposal also allows files within subpaths.
For example:
example.com/docs/llms.txt
can describe the documentation section specifically.
Step 5: Keep It Updated
An outdated llms.txt pointing to deleted documentation is not very useful.
If your website changes frequently, automation is preferable to manually editing the file every week.
How to Add llms.txt to WordPress
WordPress users have several options.
One straightforward option is Yoast SEO.
Yoast added automatic llms.txt generation to its WordPress SEO plugin. Its implementation creates the file at the root of the website and can automatically select content or allow manual selection of important pages. Yoast’s current developer documentation says the generated file is updated weekly.
You can also create the file yourself if you want complete control.
The important part is not which tool creates the file.
The quality of the information inside it matters more.
Common llms.txt Mistakes to Avoid
Turning It Into Another Sitemap
The purpose is curation.
Adding every tag, archive, author page, pagination URL, and minor blog post reduces that usefulness.
Stuffing It With Keywords
llms.txt is not a place to repeat phrases such as:
best WooCommerce plugin
best WooCommerce plugin
best WooCommerce plugin
That does not make the file more helpful.
Describe your product accurately.
Treating It as an AI Ranking Factor
There is currently no reliable evidence for making that claim, and Google explicitly says it does not use the file for rankings.
Ignoring Your Actual Website
A perfectly formatted llms.txt cannot compensate for outdated documentation, thin product pages, broken internal links, or inaccessible content.
The underlying information still matters.
Creating llms-full.txt Just Because You Can
If nobody has a practical reason to consume the full site in a single file, you may simply be creating another resource you need to maintain.
Start with the smaller problem first.
Should You Add llms.txt to Your Website?
A practical decision looks like this.
If you run a large documentation, developer, SaaS, WordPress product, or knowledge website and can generate a useful llms.txt automatically or maintain one with little effort, adding it is reasonable.
If you run a small website, it should be far down your priority list.
And if someone tells you that llms.txt is necessary to rank in Google AI results, that claim conflicts with Google’s current documentation.
A sensible priority order for most websites is:
Useful content → crawlability → site architecture → internal linking → technical SEO → structured data where relevant → accurate product/documentation content → llms.txt
not:
llms.txt → instant AI visibility.
Conclusion
llms.txt is best understood as an emerging interface between websites and AI agents.
It gives compatible agents a concise, structured guide to the information you consider most useful.
llms-full.txt takes a different approach by putting a large body of content into one file, which can be valuable for documentation ingestion, vectorization, or systems that need a large amount of context at once.
Neither file is a magic SEO technique.
Google Search explicitly says it ignores llms.txt for ranking and visibility, including its generative AI experiences.
That does not make llms.txt pointless.
For documentation-heavy websites, SaaS companies, developer platforms, and software businesses, the cost of providing a well-maintained file can be small while giving compatible AI agents a cleaner way to understand the site.
The practical rule is simple:
Create llms.txt when it provides a genuinely useful map of your content. Create llms-full.txt only when there is a real reason to provide your content in bulk. And do not let either distract you from the website and content fundamentals that already matter.
Frequently Asked Questions About llms.txt
What does llms.txt stand for?
LLM stands for Large Language Model. llms.txt is the filename proposed for providing LLM-friendly information about a website or section of a website.
Is llms.txt an official web standard?
It is better described as an emerging proposal and convention rather than a web standard comparable to established protocols such as robots.txt. The current specification is maintained at llmstxt.org and was updated to version 2 in August 2026.
Is llms-full.txt required if I use llms.txt?
No.
You can publish llms.txt without llms-full.txt.
llms-full.txt is commonly implemented by documentation platforms and sites that want to provide large amounts of documentation in a single file, but it is not required by the core llms.txt specification.
Does Google use llms.txt?
Google Search says it does not use llms.txt for Google Search, including its generative AI features.
Google also says adding the file does not positively or negatively affect search rankings or visibility.
Will llms.txt make my website appear in ChatGPT?
There is no guarantee.
An AI system has to discover the file, choose to use it, retrieve relevant resources, and decide whether those resources help answer a particular question.
Publishing llms.txt does not guarantee citations or recommendations.
Does llms.txt replace sitemap.xml?
No.
An XML sitemap helps search engines discover indexable URLs.
llms.txt is intended to provide a curated map of useful information for AI agents.
A website can use both.
Does llms.txt replace robots.txt?
No.
robots.txt communicates crawler-access instructions.
llms.txt provides information and navigation guidance.
They solve different problems.
Should I include every page in llms.txt?
Usually not.
The value of llms.txt comes largely from selecting and organizing useful resources.
Your sitemap can handle exhaustive URL discovery.
Use llms.txt to highlight authoritative product pages, documentation, guides, policies, references, or other resources an agent genuinely needs.
Should WordPress websites use llms.txt?
It depends on the site.
A WordPress software company, WooCommerce store with extensive product information, large publication, or documentation website may have a reasonable use case.
A small five-page business website probably has more important optimization work to do first.
WordPress users who already use Yoast SEO can generate and customize an llms.txt file through Yoast rather than creating it manually.
Do I need llms-full.txt for a WordPress website?
Usually not.
Consider it mainly when you maintain substantial documentation or another structured content library that an AI application may need to retrieve in bulk.
For an ordinary company website or blog, a well-curated llms.txt is usually the more sensible experiment.
Is adding llms.txt worth doing in 2026?
If implementation and maintenance are inexpensive, it can be worth adding to documentation-heavy or software-oriented sites.
Just set the expectation correctly.
Treat it as an additional machine-readable access point for compatible AI agents—not as an SEO ranking factor, traffic guarantee, or replacement for good content and technical SEO.
