Give Your AI Agent Access to E-Commerce Data with MCP

2026-03-07 · ShelfRun Team

Why AI Agents Need E-Commerce Data

AI agents are becoming the new interface for everything. Claude can browse the web, Cursor can write code, and custom agents can automate workflows. But they all share one problem: they can't easily access structured product data from e-commerce stores.

ShelfRun's MCP (Model Context Protocol) server gives any compatible AI agent 6 powerful tools for e-commerce data: scan stores, scrape products, compare prices, manage tracked stores, and more.

Set Up in Claude Desktop

// ~/.config/claude/mcp.json
{
  "mcpServers": {
    "shelfrun": {
      "command": "npx",
      "args": ["@shelfrun/mcp"],
      "env": {
        "SHELFRUN_API_KEY": "your-api-key"
      }
    }
  }
}

What Your AI Can Do

• 🔍 "Scan allbirds.com and tell me how many products they have"
• 📊 "Compare prices between Gymshark and Lululemon"
• 🔔 "Track this store and alert me when prices drop"
• 📈 "Show me the price history for this product"