Menlo Labs
Shopping MCP · v0.1.0

Product search infrastructure
for AI agents.

Shopping MCP gives agents structured access to product search, pricing, ratings, identifiers, and direct product links.

Current coverage: Amazon.com. The product interface is retailer-neutral; support for additional retailers and product sources is planned.

View tools ↘Read the human guide →
MCP endpointhttps://menlolabs.dev/mcp

From intent to product data.

One search request, normalized and cached before a provider call is considered.

User request

“Find me the best-rated Nike hoodie under $80.”
Selected toolsearch_products
Arguments
{
  "query": "Nike hoodie",
  "max_results": 3,
  "max_price": 80,
  "min_rating": 4.2,
  "sort": "best_rated"
}
Example response3 products
$64.99★ 4.7 · 9,842 reviews
$74.00★ 4.6 · 2,318 reviews
$59.95★ 4.5 · 1,104 reviews

Illustrative response shape. Live prices and availability are retrieved at request time.

Tools

Discovery for precise searches and planning for broader shopping needs.

Tool

search_products

Search the currently supported product source with a free-form query, explicit filters, and a supported sort order.

Key inputs

  • query
  • max_results
  • sort
  • min_price
  • max_price
  • min_rating
  • exclude_sponsored

Returns

  • Product name and ASIN
  • Price and currency
  • Rating and review count
  • Image and tracked product URL
  • Prime and sponsored flags
Example input
{
  "query": "Nike hoodie",
  "max_results": 3,
  "max_price": 80,
  "min_rating": 4.2,
  "sort": "best_rated"
}
Abbreviated response
{
  "status": "success",
  "cache_status": "hit",
  "products": [{
    "asin": "B0EXAMPLE",
    "name": "Nike Club Fleece Hoodie",
    "product_url": "https://www.amazon.com/dp/B0EXAMPLE",
    "tracked_url": "https://menlolabs.dev/r/a8K2mP",
    "price": { "amount": 64.99, "currency": "USD" },
    "rating": 4.7,
    "review_count": 9842
  }]
}

Tool

build_product_kit

Turn a broad shopping theme into up to three concrete categories and product searches, with a strict two-miss provider budget.

Key inputs

  • theme
  • budget
  • max_categories
  • products_per_category
  • sort

Returns

  • Normalized theme
  • Planned categories and reasons
  • Search query per category
  • Products with tracked URLs
  • Partial status when limits apply
Example input
{
  "theme": "I'm returning to recreational running",
  "budget": 250,
  "max_categories": 3,
  "products_per_category": 3
}
Abbreviated response
{
  "theme": "returning to recreational running",
  "status": "success",
  "categories": [{
    "category": "Running shoes",
    "reason": "Core footwear for regular running.",
    "search_query": "beginner daily trainer running shoes",
    "products": [/* provider-derived products */]
  }]
}

How it works

A small, inspectable path from agent intent to structured product results.

  1. 01AI agent

    Sends a constrained product request.

  2. 02Shopping MCP

    Normalizes, checks cache, and enforces limits.

  3. 03Product provider

    Retrieves current source data when needed.

  4. 04Structured results

    Returns product facts and tracked URLs.

Documentation

Read the operating instructions.

The agent guide covers connection details, tool selection, constraints, examples, limitations, and response semantics.

For Humans →For Agents →