DEV Community

BuyWhere
BuyWhere

Posted on

Build a Price-Comparing AI Shopping Agent in 10 Minutes with BuyWhere MCP

#ai #mcp #agents #tutorial

Build a Price-Comparing AI Shopping Agent in 10 Minutes

The Model Context Protocol (MCP) is how AI agents connect to the tools and data they need. In this tutorial, you will set up a BuyWhere MCP server that gives any AI agent real-time access to 296M+ products across 100K+ merchants -- ready for price comparison, deal finding, and autonomous shopping workflows.

What You Will Build

  • A working MCP server connected to a live product catalog
  • 8 tools for searching, comparing, and finding deals on real products
  • A Claude Desktop / Cursor integration that can search products by natural language

Prerequisites

  • Node.js 18+
  • A free BuyWhere API key (get one here)
  • Claude Desktop, Cursor, or any MCP-compatible client

Step 1: Get Your API Key

{"error":"agent_name is required"}

You will get back an API key instantly. The free tier includes 10,000 API calls/month.

Step 2: Install the MCP Server

changed 94 packages in 2s

33 packages are looking for funding
run npm fund for details

Step 3: Configure Your MCP Client

Add this to your Claude Desktop claude_desktop_config.json:

Step 4: Start Using the Tools

Once connected, you have access to 8 MCP tools:

Tool Description
search_products Search across 296M+ products by keyword, category, or region
get_product Get detailed info for a specific product
compare_products Side-by-side price comparison across merchants
find_best_price Find the lowest price for any product
get_deals Browse current deals and discounts
list_categories Browse available product categories
find_similar Find similar products for comparison
ingest_products Add custom products to the catalog

Example: Natural Language Price Comparison

Ask your AI agent:

"Find me the cheapest iPhone 16 Pro 256GB across all merchants in Singapore"

The agent will:

  1. Call search_products with the query and region SG
  2. Call compare_products on the top results
  3. Return a ranked list with prices, merchants, and links

Example: Deal Discovery

"What are today's best deals on wireless earbuds under $50 in the US?"

The agent uses get_deals filtered by category and price range, then find_best_price to verify.

Supported Regions

BuyWhere covers 9 regions: US, Singapore, Japan, Vietnam, Thailand, Indonesia, Philippines, Malaysia, India

Why MCP for Shopping?

Traditional product APIs require one integration per merchant, manual data refresh, and a human UI. With MCP + BuyWhere:

  • One API, all merchants -- query 100K+ stores in a single call
  • Real-time data -- prices update continuously
  • Agent-native -- built for AI, not humans browsing storefronts

Resources

BuyWhere is the product catalog API built for AI agents. Free tier available at buywhere.ai.

Top comments (0)