DEV Community

BuyWhere
BuyWhere

Posted on

Build a Price Comparison Agent with BuyWhere + LlamaIndex

#llamaindex #ai #mcp #ecommerce

Build a Price Comparison Agent with BuyWhere + LlamaIndex

If you're building AI agents with LlamaIndex, your agents can now query BuyWhere for structured price comparisons across 9 countries and get back JSON that's ready for downstream processing.

Use case: An agent that helps users find the best deal on a laptop, then routes them to the retailer with checkout context pre-filled.

The Integration

The integration is simple — wrap BuyWhere's search endpoint as a LlamaIndex FunctionTool:

from llama_index.core.tools import FunctionTool
import requests

def buywhere_search(product: str) -> dict:
    response = requests.get(
        f"https://buywhere.ai/api/search?q={product}"
    )
    return response.json()

price_tool = FunctionTool.from_defaults(
    fn=buywhere_search,
    description="Search BuyWhere for product prices across 9 countries"
)
Enter fullscreen mode Exit fullscreen mode

That's it — your LlamaIndex agent can now search for prices, compare across markets, and make recommendations backed by real-time data.

What BuyWhere Returns

Each search returns a standardized product schema:

{
  "title": "MacBook Air M4 13-inch",
  "brand": "Apple",
  "price": 1299.00,
  "currency": "SGD",
  "merchant": "Lazada",
  "country": "SG",
  "in_stock": true,
  "url": "https://..."
}
Enter fullscreen mode Exit fullscreen mode

Your agent gets structured data it can reason about — no HTML scraping, no regex parsing.

Why This Matters for Agentic Commerce

The missing piece for most AI shopping agents is reliable, structured product data. BuyWhere provides:

  • 11M+ products across Singapore, Southeast Asia, and US markets
  • Standardized schema — title, brand, GTIN, price, currency, stock, merchant
  • Sub-second latency — your agent doesn't wait
  • Free tier — 1,000 API calls/month, no credit card

Getting Started

  1. Get an API key at buywhere.ai/api-keys
  2. Install LlamaIndex: pip install llama-index
  3. Wrap the API call as shown above
  4. Add it to your agent's tool list

Or use the MCP server directly:

npx @buywhere/mcp-server --api-key YOUR_KEY
Enter fullscreen mode Exit fullscreen mode

What's Next

We're seeing early adopters build:

  • Deal-finding agents that compare prices across 3+ retailers in real-time
  • Procurement copilots that source the cheapest supplier for B2B orders
  • Price monitoring agents that alert when a target price is hit

If you're building something similar, we'd love to hear about it.

BuyWhere is an open-source price comparison API for AI agents. GitHub | Docs

Top comments (0)