/ llmtxt.info
Français

llms.txt for Gatsby

Two approaches for Gatsby: a zero-config static file in static/, or programmatic generation in gatsby-node.js using the GraphQL data layer to build links from your content.

Last updated:

Option 1: static file in static/

Gatsby copies everything in the static/ directory directly into the public/ build output. Place your file at static/llms.txt and it will be available at /llms.txt after gatsby build.

static/llms.txt — directory structure 
# Gatsby static file approach
#
# Place your file at: static/llms.txt
# Gatsby copies everything in static/ directly to the public/ build output.
#
# Project structure:
# your-gatsby-site/
# ├── static/
# │   └── llms.txt   ← add this
# ├── src/
# └── gatsby-config.js

This approach requires no code changes and works with all Gatsby deployment targets: Netlify, Vercel, Cloudflare Pages, AWS Amplify, and self-hosted.

Option 2: generate in gatsby-node.js

For sites where documentation pages are sourced from Markdown, MDX, or a CMS via the Gatsby data layer, use the onPostBuild lifecycle hook in gatsby-node.js to generate llms.txt after the build completes. This ensures the link list is always in sync with your content.

gatsby-node.js 
// gatsby-node.js
// Generate llms.txt from your GraphQL data layer
const { writeFileSync } = require('fs');
const { resolve } = require('path');

exports.onPostBuild = async ({ graphql }) => {
  // Query your documentation pages (adjust the query to your data model)
  const result = await graphql(`
    query {
      allMarkdownRemark(
        filter: { frontmatter: { collection: { eq: "docs" } } }
        sort: { frontmatter: { order: ASC } }
      ) {
        nodes {
          frontmatter {
            title
            description
          }
          fields {
            slug
          }
        }
      }
    }
  `);

  if (result.errors) {
    throw result.errors;
  }

  const docs = result.data.allMarkdownRemark.nodes;
  const siteUrl = 'https://yoursite.com';

  const links = docs
    .map(
      (doc) =>
        `- [${doc.frontmatter.title}](${siteUrl}/docs/${doc.fields.slug}/): ${doc.frontmatter.description ?? ''}`
    )
    .join('\n');

  const content = [
    '# Your Site',
    '',
    '> One-sentence description of your project.',
    '',
    '## Documentation',
    '',
    links,
    '',
    '## Optional',
    '',
    `- [Changelog](${siteUrl}/changelog/): Release history.`,
  ].join('\n');

  // Write to public/ (Gatsby's build output directory)
  writeFileSync(resolve(__dirname, 'public', 'llms.txt'), content, 'utf-8');
};

The onPostBuild hook receives the same graphql function used in page queries. Adjust the query to match your data model — the example uses allMarkdownRemark but you can query any Gatsby source plugin (allMdx, allContentfulPage, etc.).

Gatsby plugin approach

If you prefer a plug-and-play solution, several community plugins generate llms.txt automatically. Search npm for gatsby-plugin-llms-txt. Alternatively, the approach above in gatsby-node.js is straightforward enough that a custom plugin adds little value for most projects.

You can also use gatsby-plugin-sitemap as a reference — it uses the same onPostBuild pattern to write an XML file to public/.

Verify

Build and verify 
# Build
gatsby build

# Check generated file
cat public/llms.txt | head -10

# After deploying, verify the live URL
curl -I https://yoursite.com/llms.txt
# Expected: content-type: text/plain; charset=utf-8

Related guides

Sources