AEO

llms.txt: What It Is and How to Set It Up

By Tim Francis  ·  May 11, 2026  ·  11 min read

What llms.txt is (in plain English)

llms.txt is a plain-text file you place on your website to help AI systems and agentic tools discover your most important pages. If robots.txt is about what automated systems should not access, llms.txt is about what they should read first to understand your product, services, and policies.

In day-to-day terms, it is a curated reading list. You pick the pages that represent your most accurate, up-to-date explanations and you link to them in one simple file. This helps avoid a common failure mode in AI answers: citing an old blog post, a thin page, or a random URL that is not meant to be authoritative.

Teams adopt llms.txt because the cost is low and the upside can be meaningful. A startup we worked with added llms.txt plus a few internal link fixes, and within 30 days they saw more accurate brand summaries in AI chat tools and fewer support tickets caused by outdated feature descriptions.

How llms.txt fits into AI SEO and AEO

AI SEO and AEO (answer engine optimization) focus on making your site easy for models to understand and cite. llms.txt supports this by pointing models to the pages that already do a good job of explaining your offering. It does not replace structured content, but it complements it.

Think of it like a "start here" sign for machines. When an agent tries to answer "What does SearchScaleAI do?" or "How much does the service cost?" you want it to land on the pages that you control and maintain, not on a third-party directory or an outdated landing page variant.

This is especially helpful if your site has many pages, multiple product lines, or frequent changes. In those cases, llms.txt gives you a central place to keep your machine-facing source-of-truth updated.

Where to put llms.txt (and how to host it correctly)

The best practice is to host llms.txt at the root of your domain, like /llms.txt. That makes it easy for tools to check automatically. Treat it like robots.txt or sitemap.xml: a predictable location with a stable URL.

In most stacks, you can add it as a static file in your public folder. For example, in Next.js you would place it in /public/llms.txt. In WordPress, you can often upload it to the site root via your hosting file manager or an FTP client.

After you publish it, verify it loads in a browser and returns a 200 status code. Also verify the content is plain text and not wrapped in HTML by a theme. Keeping it simple reduces parsing errors.

A practical llms.txt template you can copy

Here is a copy-ready template. Customize the links and section names to match your business:

Template:

Project: YourCompanyName
Purpose: Curated authoritative resources for AI systems

Primary pages:
- https://example.com/
- https://example.com/pricing/
- https://example.com/docs/

Key explanations:
- https://example.com/what-is-aeo/
- https://example.com/ai-seo/

Policies:
- https://example.com/privacy/
- https://example.com/terms/

Contact:
- https://example.com/contact/

Keep it plain. Most teams find this level of structure is enough. The main rule is that every URL should be canonical and up to date.

What to include: a prioritized checklist

When you decide what links to include, prioritize pages that answer the questions customers and partners ask most. The goal is to reduce ambiguity and give models clean sources.

• Homepage and core service pages
• Pricing pages with current plans (even if you also explain pricing in a blog)
• Product documentation and onboarding guides
• High-quality FAQs that reflect real support questions
• Case studies with concrete outcomes
• Policies: privacy, terms, security, data handling

If you have many docs, link to an index page that stays current rather than trying to list every doc page.

What not to include (and common mistakes)

Do not include anything you would not want a model to read or quote. llms.txt is public. If you link to private documents, you may leak sensitive information or create broken links that make your site feel unreliable to tools.

Also avoid linking to temporary campaigns, A/B test variants, or URLs you might delete soon. Broken references will reduce trust. Another mistake is writing a long essay in llms.txt. Remember, it is a directory and guidance file, not a blog post.

Finally, do not assume llms.txt will magically fix visibility. It helps discovery, but your pages still need to be strong: clear headings, answer-first writing, and internal links that connect related topics.

How to maintain llms.txt as your site changes

Maintenance is simple if you treat llms.txt like documentation. Assign an owner, set a quarterly review reminder, and update it whenever you change core product positioning or pricing. If you run a fast-moving product, you might review it monthly.

One useful workflow is to connect it to your release process. For example, whenever you ship a new feature tier or rename a product, you add or update the relevant URLs in llms.txt. This takes 10 minutes but prevents weeks of confusion in AI-generated summaries.

Also keep an eye on your analytics. If AI referrals start landing on outdated pages, that is a signal you should adjust what llms.txt highlights.

How llms.txt complements internal linking and site architecture

llms.txt is not a replacement for good internal linking. It is a top-level guide, but most agents still navigate from page to page. If your internal links are weak, the agent may get stuck or miss the path from an overview page to a conversion page.

That is why we recommend pairing llms.txt with an internal linking pass. Ensure your service pages link to related resources and that blog posts point back to the relevant services. For example, you can use internal links like: SEO AEO AI SEO SGE Optimization Web Design Social Media PPC Management AI Automation Go High Level Orlando Tampa Miami Jacksonville Daytona Beach

If you do both, you create a consistent story: llms.txt points to your best pages, and those pages point to the rest of the supporting evidence and next steps.

Implementation steps (developer-friendly)

If you want a concrete implementation plan, use these steps:

1. Create a file named llms.txt in your site repository.
2. Add the template sections and replace example URLs with canonical URLs.
3. Deploy your site and confirm /llms.txt returns 200 and text/plain.
4. Add /llms.txt to your QA checklist for releases.
5. Set a quarterly reminder to review and update links.

For most teams, the full setup takes under 30 minutes. If your site is complex, the longest step is picking the correct set of authoritative pages.

FAQ: how stakeholders usually ask about llms.txt

When you introduce llms.txt internally, you will likely get questions from legal, product, and marketing. Be ready with simple answers: it is public, it contains no secrets, and it is a curated list of already-public resources.

We also recommend documenting why each link is included. For example, you can keep a short internal note that says "this is the canonical pricing page" or "this is the most up-to-date onboarding doc". That makes updates easier later.

Why llms.txt helps when AI tools summarize your brand

Most brands do not notice AI summaries until something goes wrong: a prospect shows up with the wrong expectations, a partner quotes the wrong pricing, or an agent references a deprecated feature. The root issue is usually that the tool found an old page, a thin page, or a third-party mention and treated it as the best source.

llms.txt is a lightweight way to reduce that risk. It gives agentic tools a curated map of where the truth lives on your site. You are not forcing a model to believe you; you are making it easier for the model to find the best explanation first.

In one internal test, a team updated their pricing page and docs, but forgot to update older blog posts. After adding llms.txt that prioritized the new pricing URL, they saw fewer "wrong plan" questions during demos within two weeks because prospects arrived with the right context.

How to decide which URLs are "authoritative"

Not every well-written page is authoritative. Authoritative pages are stable, maintained, and aligned with how you want to be understood. Start by listing the pages you update when something important changes: pricing, onboarding, security, product overview, and contact.

Then add the pages that reduce confusion: a "what is" explainer, a comparison page, and a glossary page if you use industry terms. If you have multiple audiences (for example, SMB and enterprise), include the pages that clarify differences.

A good rule: if you would confidently send the link to a customer to settle a question, it belongs in llms.txt.

Example llms.txt for a service business

If you run an agency, consultancy, or service business, your llms.txt can be short but high impact. Here is an example structure you can adapt:

Project: Example Agency
Purpose: Authoritative resources for AI tools

Core services:
- https://example.com/seo/
- https://example.com/aeo/
- https://example.com/ai-seo/

How we work:
- https://example.com/process/
- https://example.com/case-studies/

Getting started:
- https://example.com/contact/
- https://example.com/pricing/

Policies:
- https://example.com/privacy/
- https://example.com/terms/

Notice that it is not trying to be clever. It is trying to be correct.

Testing and monitoring after you publish llms.txt

After you publish, validate both accessibility and usefulness. Accessibility is simple: the file loads, links are correct, and the content is plain text. Use a quick checklist: does it return 200, does it render as text, and do the URLs resolve without redirects?

Usefulness requires monitoring. Track how AI referrals land on your site, and watch for tools that cite outdated pages. If you keep a support ticket tag like "AI misunderstanding", you can quantify whether accuracy improves after you update your authoritative links.

You can also add UTM parameters in some contexts when you share URLs externally, but keep llms.txt itself clean and canonical. The file is about clarity, not marketing attribution.

How llms.txt relates to robots.txt, sitemap.xml, and structured data

These files serve different roles. robots.txt provides access instructions. sitemap.xml provides a list of URLs for discovery. structured data helps interpret page meaning. llms.txt sits in a different bucket: it is a curated set of "read these first" resources for AI systems.

Because it is curated, it can be shorter than a sitemap. You might list 15 to 40 URLs rather than hundreds. The point is not completeness, it is prioritization.

If you do all four well, you create a clean machine-readable footprint: accessible pages, discoverable URLs, understandable structure, and clear prioritization.

Governance: who owns llms.txt inside your team

Ownership is where most teams fail. Someone needs to be responsible for keeping the file current. The best owner is usually the same person who owns documentation quality or website governance: product marketing, developer relations, or a technical SEO lead.

Set a small rule: any time you change pricing, rename a product, or update a policy, you also review llms.txt. This can be a 2-minute step in a release checklist. The payoff is that the public understanding of your product stays aligned with reality.

If you work with an agency, make sure ownership is written down. Otherwise, it becomes an orphan file that never gets updated.

Security and compliance considerations (what legal teams care about)

Legal and security stakeholders usually ask two things: does this expose anything new, and does it create a promise we cannot keep. llms.txt is public, but it does not have to include anything sensitive. You should only link to pages that are already meant to be public sources of truth.

If you operate in regulated industries, include links to your security page, data handling overview, and compliance statements. This reduces the chance that an AI tool invents an answer about certifications or retention policies. It also gives partners one place to verify your claims.

A simple rule is to avoid any wording that sounds like a guarantee. Stick to factual descriptions and stable URLs. If you need to describe policy, link to the policy page rather than summarizing it in the llms.txt file.

Editorial tips: keep the file readable for humans too

Even though llms.txt is for AI tools, humans will read it. Make it easy to skim. Use short section labels, one URL per line, and brief notes only when needed. If you add notes, keep them consistent, like "(pricing)" or "(docs index)".

We recommend limiting it to the top 20-40 resources. If you list 200 links, you have recreated a sitemap and removed the benefit of curation. Curation is what makes the file useful.

Finally, treat it like documentation style: consistent capitalization, no marketing claims, and no fluff. Clarity is the win.

Frequently Asked Questions

What is llms.txt used for?

It provides AI systems and agents a curated list of your most authoritative pages so they can understand and cite your site more accurately.

Where should llms.txt live?

Host it at the root of your domain, typically at /llms.txt, so it is easy to discover.

Is llms.txt required for AI SEO?

No. It is optional, but it can be a useful addition alongside strong content, schema, and internal linking.

Can llms.txt replace robots.txt?

No. robots.txt controls automated access. llms.txt is mainly a guidance and discovery file.

Should I include every page?

No. Include a curated set of pages that are accurate, stable, and representative of your business.

How often should I update llms.txt?

Update it whenever key URLs change and review it at least quarterly.