@lg2/storyblok-llms-nuxt

A Nuxt 3 module that automatically generates llms.txt files from Storyblok CMS content during the build process. This module fetches content from Storyblok, extracts a specified field, and writes it to a public file that can be used for LLM training, documentation, or content discovery.

Features

• 🔄 Automatically fetches content from Storyblok CMS during build
• 📝 Extracts and writes content to llms.txt (or custom path)
• 🔍 Flexible field extraction from various Storyblok content structures
• ⚠️ Graceful error handling with development mode support

Installation

Install the package using your preferred package manager:

# npm
npm install @lg2/storyblok-llms-nuxt

# yarn
yarn add @lg2/storyblok-llms-nuxt

# pnpm
pnpm add @lg2/storyblok-llms-nuxt

Then add it to your nuxt.config.js:

export default defineNuxtConfig({
  modules: [
    '@lg2/storyblok-llms-nuxt',
  ],
  storyblokLlms: {
    accessToken: process.env.STORYBLOK_ACCESS_TOKEN,
    // ... other options
  },
})

Configuration

The module accepts the following configuration options:

| Option | Type | Default | Description | |--------|------|---------|-------------| | accessToken | string | Required | Your Storyblok access token (from environment variable) | | outputPath | string | 'public/llms.txt' | Relative path where the output file will be written | | fieldName | string | 'llms' | The field name to extract from Storyblok content | | modelName | string | 'common' | The Storyblok story slug/model name to fetch | | language | string | 'en' | The language code to fetch from Storyblok |

Example Configuration

export default defineNuxtConfig({
  storyblokLlms: {
    accessToken: process.env.STORYBLOK_ACCESS_TOKEN,
    outputPath: 'public/llms.txt',
    fieldName: 'llms',
    modelName: 'config',
    language: 'en',
  },
})

How It Works

1. Build Hook: The module hooks into Nuxt's build:before and Nitro's nitro:build:before hooks
2. API Fetch: Fetches content from Storyblok's CDN API using the configured story/model
3. Field Extraction: Searches for the specified field in multiple possible locations:
   • Direct field: content.llms
   • Component body: content.body[].llms
   • Configuration object: content.configuration.llms
4. Content Handling: Supports various content formats:
   • Plain strings
   • Markdown fields (content, text, or markdown properties)
5. File Writing: Writes the extracted content to the specified output path

Storyblok Content Structure

The module looks for the llms field (or your custom fieldName) in the following locations:

Option 1: Direct Field

{
  "content": {
    "llms": "Your content here..."
  }
}

Option 2: In Component Body

{
  "content": {
    "body": [
      {
        "component": "some-component",
        "llms": "Your content here..."
      }
    ]
  }
}

Option 3: In Configuration

{
  "content": {
    "configuration": {
      "llms": "Your content here..."
    }
  }
}

Environment Variables

Set the following environment variable:

STORYBLOK_ACCESS_TOKEN=your-storyblok-access-token

Error Handling

• Missing Access Token: The module will log a warning and skip generation if STORYBLOK_ACCESS_TOKEN is not set
• API Errors: In development mode, errors are logged but don't fail the build. In production, errors will fail the build
• Missing Field: If the specified field is not found, a warning is logged and the build continues

Use Cases

• LLM Training: Generate training data files for language models
• Documentation: Create public documentation files from CMS content
• Content Discovery: Make site content discoverable by crawlers and AI tools
• SEO: Provide structured content for search engines and AI assistants

Requirements

• Nuxt 3.x
• Node.js >= 18.0.0

Development

The module uses Nuxt's module system and integrates with:

• @nuxt/kit for module definition
• ofetch for API requests
• Node.js fs module for file operations

Troubleshooting

File Not Generated

1. Check that STORYBLOK_ACCESS_TOKEN is set correctly
2. Verify the modelName matches your Storyblok story slug
3. Ensure the fieldName exists in your Storyblok content
4. Check the build logs for error messages

Wrong Content Extracted

1. Verify the field structure in your Storyblok content
2. Check that the field contains the expected data format
3. Review the extraction logic in the extractLLMsField function

Build Fails in Production

• Ensure your Storyblok access token has the correct permissions
• Verify the API endpoint is accessible
• Check network connectivity during build

License

MIT