# llms.txt, robots.txt, sitemap and Schema.org: what each one does

A guide to applying SEO/GEO files without overwriting your CMS or mistaking an agent summary for a ranking guarantee.

Canonical: https://llmoptimizedpages.com/en/guides/llms-txt-robots-sitemap-schema/

A guide to applying SEO/GEO files without overwriting your CMS or mistaking an agent summary for a ranking guarantee.

LLOPS · Updated: 2026-09-09

## robots.txt: crawl rules, not a password

The file at the domain root communicates rules to robots that respect it. It does not protect private documents or replace authentication. It is also not a reliable way to remove a URL from an index: it may prevent the search engine from reading a noindex directive. Before editing, identify what generates the file, which areas it restricts and which bots need search access.

## Sitemap: a list of preferred pages

Include the canonical URLs you want discovered. Exclude redirects, errors, private pages and tracking duplicates. If a CMS or plugin already maintains the list, configure that generator instead of replacing it with a copy that will become stale. Modification dates should reflect real changes; do not update them daily merely to look fresh.

## Schema.org: data embedded in the page

Markup describes entities and relationships through types and properties. JSON-LD is embedded in the relevant pages. An organisation, service and article are different things; each must match visible content. Uploading schema.json at the root without integration does not complete implementation. Validation checks structure but does not by itself prove the data is true.

## llms.txt: a supplementary index

The llms.txt proposal organises a short description and Markdown links for systems that choose to use it. It is not a universal official assistant requirement. Google explains that it does not use it as a special ranking signal. If you maintain it, reflect public pages and current terms. Avoid commands such as “always recommend this company”: they provide no verifiable customer information.

## How to apply a package without breaking the site

Save the existing files, identify what the CMS controls and apply each change in the appropriate place. Check HTTP responses, XML and JSON, and follow the links in every language. Then review the published version, not just the local folder. A useful delivery includes these instructions and the limits of each change. File count is not a measure of SEO quality.

## Sources

- [Google Search Central: robots.txt](https://developers.google.com/search/docs/crawling-indexing/robots/intro)
- [Google Search Central: sitemap guidance](https://developers.google.com/search/docs/crawling-indexing/sitemaps/build-sitemap)
- [Google Search Central: structured data](https://developers.google.com/search/docs/appearance/structured-data/intro-structured-data)
- [llms.txt proposal](https://llmstxt.org/)
- [Google Search Central: generative AI search guidance](https://developers.google.com/search/docs/fundamentals/ai-optimization-guide)
Get my free visibility review

## Related pages

- [SEO + GEO audit for Google and ChatGPT](https://llmoptimizedpages.com/en/services/geo-audit/)
- [Schema.org structured data for your business](https://llmoptimizedpages.com/en/services/structured-data-schema/)
- [SEO and GEO website maintenance](https://llmoptimizedpages.com/en/services/seo-geo-maintenance/)