Pimcore MCP Server

MCP server for Pimcore PIM (Product Information Management) integration.

Features

• get_products - Query products with filters (category, search, limit)
• get_categories - List all product categories
• get_stats - Catalog statistics (total products, languages, gaps)
• save_descriptions - Save AI-generated descriptions (with confirmation)
• save_translations - Save translations (with confirmation)
• save_seo - Save SEO metadata (with confirmation)

Installation

npm install
cp .env.example .env

Configuration

Edit .env:

# AI Integration API
API_URL=wss://api.ai-integration.com
API_TOKEN=sk_live_your_token_here

# Pimcore credentials
PIMCORE_URL=https://your-pimcore.com
PIMCORE_API_KEY=your_pimcore_api_key

# Optional
PIMCORE_DEFAULT_LANGUAGE=en

Usage

Development

npm run dev

Production

npm run build
npm start

Docker

docker build -t mcp-pimcore .
docker run --env-file .env mcp-pimcore

Tools Reference

get_products

Query products from the Pimcore catalog.

Parameters: | Name | Type | Description | |------|------|-------------| | category | string | Filter by category name | | search | string | Search in name/SKU | | limit | number | Max results (default: 10) | | offset | number | Pagination offset | | missingDescription | boolean | Only products without description |

Example:

{
  "name": "get_products",
  "arguments": {
    "category": "Footwear",
    "limit": 5
  }
}

get_categories

List all product categories.

Parameters: None

get_stats

Get catalog statistics.

Returns:

• Total products
• Products without descriptions
• Available languages
• Category breakdown

save_descriptions

Save AI-generated descriptions. Requires confirmation.

Parameters: | Name | Type | Required | Description | |------|------|----------|-------------| | items | array | Yes | [{productId, description}] | | language | string | No | Target language (default: en) | | confirmed | boolean | Yes | Must be true to save |

Example:

{
  "name": "save_descriptions",
  "arguments": {
    "items": [
      {"productId": "123", "description": "Premium leather boots..."},
      {"productId": "456", "description": "Elegant handbag..."}
    ],
    "language": "en",
    "confirmed": true
  }
}

save_translations

Save product translations. Requires confirmation.

Parameters: | Name | Type | Required | Description | |------|------|----------|-------------| | items | array | Yes | [{productId, translations: {lang: text}}] | | confirmed | boolean | Yes | Must be true to save |

save_seo

Save SEO metadata. Requires confirmation.

Parameters: | Name | Type | Required | Description | |------|------|----------|-------------| | items | array | Yes | [{productId, title, description, keywords}] | | confirmed | boolean | Yes | Must be true to save |

Pimcore API Requirements

Your Pimcore instance needs:

1. REST API enabled with API key authentication
2. Data Object classes for Products with fields:
   • name (localized)
   • description (localized)
   • sku
   • categories (relation)
3. Permissions for read/write on Product objects

Testing

npm test

Troubleshooting

Connection refused

• Check API_URL is correct
• Verify API_TOKEN is valid
• Ensure firewall allows outbound WebSocket

Pimcore API errors

• Verify PIMCORE_URL and PIMCORE_API_KEY
• Check API permissions in Pimcore admin
• Ensure REST API is enabled

Tool not found

• Restart server after tool changes
• Check tool is exported in src/tools/index.ts

License

MIT