SEO Skill

llms.txt generator

Start with a sitemap or add links yourself, choose the pages an agent should read first, then copy or download an llms.txt file.

Start from a sitemap

Enter a public sitemap URL, then choose the pages to add. You can edit every label and link before downloading the file.

We check up to five sitemap files and show up to 1,000 pages. We do not read the pages themselves.

Link sections

Group the pages an agent should open first. Keep secondary pages in an Optional section.

Generated llms.txt

Add a project name to create the file.

What this generator creates

The llms.txt proposal uses Markdown. A file starts with one H1 title and can add a short blockquote summary, notes, and H2 sections containing links. A section named Optional identifies secondary pages that can be skipped when a consumer needs a shorter context.

Read the current llms.txt proposal and format before publishing the file. This generator checks its own fields and creates that basic structure. It does not fetch the linked pages or prove that another service reads them.

llms.txt does not improve Google rankings

Google Search says it ignores llms.txt files for visibility and rankings. Publish one only when an agent or service you care about explicitly uses it. Keep normal crawling, indexing, internal links, and useful page content as separate work.

See Google's current generative AI search guidance for that distinction.

Generate a draft from a real crawl

This web tool reads URLs from the sitemap files you submit. It does not crawl the pages or decide which ones matter. If you want automatic page selection from indexable crawl evidence, install the package and use the crawl-driven llms.txt generator. Review the selected URLs before publishing either result.

npm i -g seo
seo start
seo llms generate --project example --output llms.txt

Check the published file

After publishing, run the llms.txt audit to check the response, content type, links, redirects, duplicates, and crawl coverage. The audit records those observations without turning the optional file into a search requirement.