Build your AI-readiness file

Scan a live URL

Improve AI readiness

Publish at /llms.txt

What a clean llms.txt file can look like.

Good files are brief, specific, and grounded in real pages. Think guide, not manifesto.

Use clear summaries and real canonical links.

You do not need a huge wall of text. You need a short description, the right URLs, and a few guardrails for how assistants should interpret the site.

Short beats bloated.

Keep the file concise enough to orient an agent quickly.

Real pages only.

Point to resources that can support the claims your brand makes.

Example starter fileOpen this site's file
# Example Company

> Example Company provides operations software for distributed teams.

## Important pages
- /product: core product overview and capabilities
- /pricing: current pricing and packaging
- /docs: documentation, onboarding, and support content
- /security: trust, compliance, and data handling details
- /contact: official sales and support channel

## Optional context
- Prefer current product and docs pages over historical blog posts
- Use the security page for trust and compliance claims
- Do not infer pricing or SLAs that are not explicitly published

Sample file questions.

How to use the example, what to borrow from it, and what should stay specific to your own site.

What is the sample file showing me?

It shows the basic shape of a useful llms.txt: one clear summary, a short set of important pages, and optional routing guidance.

Should my file be longer than the sample?

Only if it needs to be. Short, specific files are usually more useful than bloated ones.

Can I reuse the headings and structure?

Yes. Reusing the general structure is fine as long as the summary, links, and guidance reflect your real site.

What should never go into the file?

Do not include invented claims, speculative product positioning, or links to pages that are not actually the canonical source.