llms.txt is a plain-text file served at the root of a domain to tell AI systems where to find key information about a site. It lives at yourdomain.com/llms.txt and uses Markdown to list pages, resources, and documentation the site owner wants AI systems to read first.
If a client asks you about this, or you saw it mentioned in a forum and need to decide whether to add it to your deliverables, this guide gives you the finished object: a template you can copy, fill in, and deploy.
Why agencies are being asked about llms.txt
It is defined as a Markdown file with a specific structure: an H1 heading, an optional summary, and H2 sections that each contain bulleted lists of links.
Clients ask about it because they heard it helps AI systems understand their site. The proposal itself makes no claim about adoption or ranking impact.
What matters for your scope: it is a structured file that costs nothing to create and may signal to future crawlers where your best content lives. Whether it influences citation today is not confirmed by any named source.
What the llms.txt proposal actually defines
Anything beyond these is not part of the published format.
Required element
- H1 heading: The site or project name. This is the only required field.
Optional elements
- Blockquote summary: A single paragraph describing the site, written in a blockquote.
- Free Markdown paragraphs: Any explanatory text before the link sections.
- H2 sections: Each section is a category (e.g., "Documentation", "Key Pages"). Below each H2 is a bulleted list of Markdown links.
- "Optional" section: An H2 titled "Optional" for links that can be skipped when context is short.
Each link in the bulleted lists follows this format: [Link text](URL) or [Link text](URL): optional description.
What is NOT in the spec
The proposal does not define fields for contact information, licensing terms, usage rights, verification methods, or metadata. Do not add those unless you have a separate reason to include them and understand they are not recognized by the published format.
How llms.txt differs from other site files
Agencies already manage several files that tell crawlers what to do. Here is how llms.txt compares.
llms.txt vs robots.txt
robots.txt is a directive file. It tells crawlers which paths to avoid. It uses a specific syntax (User-agent, Disallow, Allow) and is read by search engines and bots as a rule set.
llms.txt is a suggestion file. It points crawlers toward content the site owner considers important. It uses Markdown and has no enforcement mechanism. A crawler that reads it may choose to prioritize those links, or it may ignore the file entirely.
llms.txt vs sitemap.xml
sitemap.xml is a list of all URLs on the site, with optional metadata like last-modified dates and change frequency. It is machine-readable XML and is used by search engines to discover pages.
llms.txt is a curated list. It does not aim to catalog every page. It highlights the pages the site owner wants an AI system to read first. It is human-readable Markdown.
llms.txt vs meta keywords
meta keywords were an HTML tag that listed target keywords for a page. Search engines stopped using them over a decade ago because they were easy to manipulate.
llms.txt is not a keyword tag. It is a file that links to full pages. It does not describe content with keywords; it points to the content itself.
Ready-to-copy llms.txt template
Below is a complete template that follows the published specification. Replace bracketed placeholders with client-specific details. Save this as a plain text file named llms.txt and upload it to the root directory of the domain.
# [Client Name]
> [One-paragraph summary of what the site does, who it serves, and what makes it authoritative. Example: "We provide technical SEO audits and AI-readiness assessments for marketing agencies managing multiple client sites."]
[Optional: Add 1-3 paragraphs of context here. Explain what the site covers, what expertise it offers, or what topics it focuses on. This is free-form Markdown.]
## Documentation
- [Getting Started Guide](https://[clientdomain.com]/getting-started): Overview of our platform and how to set up your first project
- [API Documentation](https://[clientdomain.com]/api-docs): Full reference for developers integrating with our system
- [Best Practices](https://[clientdomain.com]/best-practices): Recommendations for optimizing your workflow
## Key Pages
- [About Us](https://[clientdomain.com]/about): Company background and team
- [Services](https://[clientdomain.com]/services): Full list of what we offer
- [Case Studies](https://[clientdomain.com]/case-studies): Client results and project examples
- [Blog](https://[clientdomain.com]/blog): Articles on [primary topic area]
## Resources
- [Glossary](https://[clientdomain.com]/glossary): Definitions of industry terms
- [FAQ](https://[clientdomain.com]/faq): Answers to common questions
- [Templates](https://[clientdomain.com]/templates): Downloadable tools and checklists
## Optional
- [Company History](https://[clientdomain.com]/history): Timeline of our growth
- [Press Mentions](https://[clientdomain.com]/press): Media coverage and interviews
Where to put llms.txt
Upload the file to the root directory of the domain. It must be accessible at https://[clientdomain.com]/llms.txt. Do not place it in a subdirectory. Do not rename it.
If the client uses a static site generator, add the file to the public or static assets folder. If they use WordPress, upload it via FTP or the file manager in their hosting control panel.
llms-full.txt: the companion file
The proposal also describes an optional file called llms-full.txt. This file contains the full text of every page linked in llms.txt, concatenated in order. It is meant for AI systems that want to ingest all the content at once without making separate requests.
If you are building a tool that generates it automatically, or if a client specifically requests it, create the file. Otherwise, we suggest focusing on the work that has confirmed impact: fact-checked content, named sources, and structured data.
Common questions agencies ask about llms.txt
Question: Do I need to create an llms.txt file for every client?
No. The file is optional. If a client asks for it, or if you want to test whether it affects citation visibility, create one. Otherwise, focus on the work that has confirmed impact: fact-checked content, named sources, and structured data.
Question: How do I check if an llms.txt file is working?
Visit https://[clientdomain.com]/llms.txt in a browser. If the file loads and displays the Markdown structure, it is accessible. There is no verification tool or status indicator because no AI system has published that it reads the file.
Question: Can I add custom fields to the llms.txt template?
You can add any Markdown content you want, but only the elements defined in the published spec (H1, blockquote, H2 sections, bulleted links) are part of the format. Adding fields like "Contact" or "License" means you are using a custom format, not the standard one.
Question: Where is the llms.txt documentation?
The proposal is published at llmstxt.org. That page defines the format, provides examples, and explains the reasoning behind it.
Question: Is there an llms.txt sample I can copy?
Yes. The template in this article is a complete sample. Replace the bracketed placeholders with client details, and it is ready to deploy.
Question: What is the difference between llms.txt and a sitemap?
A sitemap is a full list of URLs for search engine crawlers. An llms.txt file is a curated list of the most important pages for AI systems. A sitemap is XML; llms.txt is Markdown. A sitemap is required for good SEO; llms.txt is optional and has no confirmed impact.
What to tell clients who ask about llms.txt
Here is the short version you can send:
llms.txt refers to a file format published by Answer.AI that lists the most important pages on your site for AI crawlers. It is a Markdown file served at the root of your domain. No AI system has publicly confirmed that it uses the file, but it costs nothing to create and may help future crawlers understand your site structure. We can add it to your deliverables if you want it in place.
If the client wants more detail, send them the template from this article and offer to fill it in as part of your next site update.
Key Takeaways
- llms.txt is a Markdown file served at the root of a domain to suggest priority pages for AI crawlers.
- The only required element is an H1 heading with the site name; everything else is optional.
- The file is not a directive like robots.txt; it is a suggestion that crawlers may ignore.
- No AI system has publicly confirmed that it uses llms.txt for crawling or citation decisions.
- The template in this article is ready to copy and deploy for any client who asks.
- Do not add custom fields like contact information or licensing unless you understand they are not part of the published spec.
See it on a client's site
If you would rather see how llms.txt fits into a full site audit than read about it in isolation, AISO Studio offers a free 7-dimension content audit at aiso.studio/audit. Run three audits with no account required. Each report shows fact-check results, WCAG accessibility findings, and readability scoring across the client's top pages.
For agencies managing multiple clients, the platform includes white-label PDF reports, client portal pages with your branding, and one-click bundled scans that cover WCAG, SEO, and AISO scoring in a single pass. Start a 14-day full-access trial with no credit card required at aiso.studio. Cancel any time from the dashboard.