---
title: "Publish an llms.txt file"
description: "llms.txt is a proposed convention: a Markdown file at the site root that maps the most useful pages for AI systems. It is not a Google standard and adoption is unproven, so publish it as a low-cost experiment, keep it current, and measure referrals instead of assuming impact."
url: "https://seo-mcp.io/t/publish-an-llms-txt-file"
---

# Publish an llms.txt file

llms.txt is a proposed convention: a Markdown file at the site root that maps the most useful pages for AI systems. It is not a Google standard and adoption is unproven, so publish it as a low-cost experiment, keep it current, and measure referrals instead of assuming impact.

## Prerequisites

- A list of canonical, indexable key pages

## Steps

1. **Select the pages worth mapping** *(general best practice)* — Choose canonical documentation, product and reference pages that answer common questions. Exclude gated, duplicate and low-value URLs.
   Check: The page list contains only URLs that return 200, are indexable and are canonical.
2. **Write the file** *(general best practice)* — Create /llms.txt in Markdown with an H1 title, a summary paragraph and sections of links with one-line descriptions. Keep it plain text.
   Check: curl your production domain/llms.txt returns 200 with text/plain or text/markdown and contains the title and at least one link.
3. **Keep it current and measure** *(general best practice)* — Regenerate it when key pages change, and track AI referrers separately to judge whether the file has any effect.
   Check: The file lists no URL that returns 4xx or 3xx, and AI referral traffic is tracked in a dedicated channel group.

## Source

@seo-mcp on 2026-10-19: https://llmstxt.org/

Canonical: https://seo-mcp.io/t/publish-an-llms-txt-file