#llmsmap.me

Guide · practical guide

How to create llms.txt: structure, examples, and validation

A step-by-step guide to publishing llms.txt for SaaS, ecommerce, documentation, and content sites.

·10 min read
How to create llms.txt: structure, examples, and validation
llmsmap.me editorial teamGuide · 10 min read

A good llms.txt file is short, stable, and intentionally curated. It should help an AI agent understand the website before it reads deeper content. Treat it as an editorial map, not a raw export of every URL.

1. Start with the project identity

Use a top-level heading for the product or organization name. Add one short paragraph that explains what the site does, who it serves, and what kind of information is authoritative.

# Example Cloud
> Developer tools for deploying and monitoring production web apps.

2. Group links by intent

Avoid a flat list of random pages. Group links into sections such as Product, Documentation, API, Pricing, Support, Policies, or Changelog. The section names should match how a user or agent would reason about the site.

## Documentation
- [Quickstart](https://example.com/docs/quickstart)
- [API reference](https://example.com/docs/api)
- [Deployment guide](https://example.com/docs/deployments)

3. Prefer canonical pages

Link to the source of truth. If a marketing page summarizes a feature but the documentation has the exact behavior, include the documentation URL. If a policy has regional variants, link to the version that applies globally or explain the difference.

4. Add llms-full.txt for dense content

llms.txt should stay compact. For long documentation, product catalogs, or knowledge-base exports, create /llms-full.txt. This lets agents decide whether they need the short overview or the full context.

5. Validate before publishing

  • The file returns HTTP 200.
  • The content is plain text or Markdown-like text.
  • Important links are absolute and canonical.
  • The file does not expose private, draft, or internal URLs.
  • The content can be understood without executing JavaScript.

Example structure

# Example Cloud
> Developer tools for deploying and monitoring production web apps.

## Product
- [Overview](https://example.com/product)
- [Pricing](https://example.com/pricing)

## Documentation
- [Quickstart](https://example.com/docs/quickstart)
- [API reference](https://example.com/docs/api)

## Support
- [Status page](https://status.example.com)
- [Contact support](https://example.com/support)

## Optional
- [Full context](https://example.com/llms-full.txt)

Once the file is live, run an audit and submit it to the llmsmap.me global directory.

Next step

Measure how ready your website is for AI search

The audit checks llms.txt, crawler access, structure, performance, and the other signals that shape AI readiness.

How to create llms.txt: structure, examples, and validation — llmsmap.me