{
  "id": 21241050,
  "title": "I put my Laravel store inside AI assistants. Here's what I learned (and the free package)",
  "url": "https://urgent.news/2026/11/22/i-put-my-laravel-store-inside-ai-assistants-heres-what-i-learned-and",
  "topic": "ai",
  "section": "AI",
  "published": "2026-11-22T16:18:33.000Z",
  "source": {
    "name": "Dev.to",
    "slug": "dev-to",
    "url": "https://dev.to/trongio/i-put-my-laravel-store-inside-ai-assistants-heres-what-i-learned-and-the-free-package-n0d"
  },
  "original_language": "en",
  "account": "I recently integrated an MCP server into my Laravel e-commerce platform, allowing AI assistants to search and interact with my live catalogue. This post details my experience, lessons learned, and the free package I built from this integration. Check out a short demo video showing how shoppers can ask for budget cameras, compare products, and add them to a cart that opens on my actual store.\n\nMCP provides an API for assistants to call tools that interact with your store's data. For a store, the most useful tools turned out to be:\n\n1. search_products: Allows assistants to search for products by words, model numbers, category, brand, price range, and in-stock status.\n2. get_product: Returns detailed information about a single product, including specs, stock, warranty, and documents.\n3. compare_products: Provides a side-by-side table comparison of multiple products.\n4. cart tools: Enable the assistant to add, update, remove items from a cart, view delivery options, and access store information (branches, hours, contacts).\n\nThe key insights from this project are:\n\n1. Present answers in a format similar to your product pages, not just raw database information. This helps build trust with shoppers when they click through to your site for more details.\n2. Always use the actual prices from your storefront for product information. Assistants should not invent prices, as this could lead to confusion or mistrust.\n3. Keep the assistant's functionality read-only until the shopper decides to add items to a cart. The search and product tools should not modify anything, and the cart tools should only create an anonymous cart with an expiry, ensuring no payment information leaves your site until the shopper completes the purchase.\n\nTo integrate this system, you'll need to implement a contract that defines how your store can be queried by the assistant. The free version of Storewire Lite provides a simple interface with three main methods:\n\n```php\nuse Hackerman\\StorewireLite\\Contracts\\CatalogRepository;\n\ninterface CatalogRepository {\n// Returns a product ID, URL, or slug based on a reference, locale, or SKU\npublic function resolve(string $reference, string $locale): ?string;\n\n// Returns detailed product information as it appears on your site\npublic function product(string $id, string $locale): ?ProductDetail;\n\n// Searches for products based on a query\npublic function search(SearchQuery $query): SearchResult;\n}\n```\n\nYou can bind your implementation of this interface to the Storewire Lite package, enabling your store to be queried by any assistant that supports MCP servers. A demo store is available at https://demo.plugins.hackerman.ge/mcp, containing 61 products for testing purposes.\n\nStorewire Lite is available as an open-source package under the MIT license on GitHub. The free version includes search_products and get_product functionality. The full Storewire package, offering additional features like compare_products, cart handoff, delivery quotes, and store info, is currently 40% off from 27 Nov to 2 Dec at plugins.hackerman.ge.\n\nAs AI assistants become more prevalent in shopping, it's essential to consider how your store's catalogue is exposed to these tools. Are you already providing this functionality, or would you consider implementing something similar? I'm open to feedback on how others are approaching this integration and welcome any questions in the comments.",
  "summary": "Shoppers have started asking AI assistants what to buy. \"I need four outdoor cameras for a small shop, around 1500 lari.\" The assistant answers something. The question is whether that answer comes from your store, with your real prices and stock, or from a guess. I run the platform behind a security-equipment store in Georgia, built on Laravel. Recently I put an MCP server on top of it, using the…",
  "key_points": [
    "Integrate MCP server into Laravel e-commerce platform",
    "Develop free package with searchproducts, getproduct, compareproducts, and cart tools",
    "Storewire Lite offers read-only functionality for AI assistants"
  ],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}