Skip to content

Llms Txt Generate MCP tool

MCP findutils:llms_txt_generate

Build an llms.txt file (llmstxt.org format): an H1 site name, a blockquote summary, optional prose, H2 sections of "- [title](url): description" links, and an "## Optional" section last. Root-relative URLs are resolved against the origin; off-origin, duplicate and placeholder links are reported as warnings. Also returns a robots.txt snippet allowing the major AI crawlers, a Link header and an HTML <link> pointing at /llms.txt. Generates text only and fetches nothing, so it cannot check that the linked pages exist.

Arguments

application/json
  • site_name

    string required

    Site or project name, used as the single H1.

  • origin

    string required

    Site origin, e.g. "https://docs.example.com". Root-relative link URLs resolve against it.

  • summary

    string required

    One or two sentences for the blockquote under the H1 (aim for 320 characters or fewer).

  • details

    string optional

    Optional plain Markdown prose after the summary (no headings).

  • sections

    array optional

    H2 sections: objects { heading, links: [{ title, url, description }] }. A section named "Optional" is moved to the end. Empty sections are dropped.

  • optional_links

    array optional

    Links for the "## Optional" section, which agents may skip: [{ title, url, description }].

  • content_signal

    string optional

    Optional Content-Signal value for the robots snippet, e.g. "search=yes, ai-input=yes, ai-train=no".

  • sitemap_url

    string optional

    Optional absolute sitemap URL added to the robots snippet.

Example arguments

Verified
{
  "site_name": "Example",
  "origin": "https://example.com",
  "summary": "Example builds small tools for developers.",
  "sections": [
    {
      "heading": "Docs",
      "links": [
        {
          "title": "Getting started",
          "url": "https://example.com/docs/",
          "description": "Install and first steps."
        }
      ]
    }
  ]
}