Amazon Product Search MCP Server

The site is live here ! : https://shopassist-sharavana.streamlit.app/

An MCP (Model Context Protocol) server that provides AI-powered Amazon product search and recommendations using FastMCP.

Features

• 🔍 Smart product search with Amazon API

• 🤖 AI-powered product recommendations using Hugging Face

• 💰 Price range filtering

• 📋 Feature-based matching

• 🎯 Tailored recommendations for Small/Medium Enterprises

Installation

1. Clone this repository and navigate to the project directory

2. Install dependencies:

   # Using uv (recommended)
   uv sync
   
   # Or using pip
   pip install -r requirements.txt

Server Setup

Running the MCP Server

# Activate your virtual environment
source .venv/bin/activate  # On Windows: .venv\Scripts\activate

# Run the server
python main.py

The server exposes one main tool:

• getdata: Search Amazon products with AI recommendations

Client Options

We provide multiple client implementations to interact with your MCP server:

1. Python Interactive Client (client.py)

A full-featured Python client with examples and interactive mode.

python client.py

Features:

• Pre-built examples (laptops, smartphones)

• Interactive search mode

• Real-time communication with MCP server

2. Command Line Interface (cli_client.py)

Quick command-line searches for automation and scripting.

# Basic search
python cli_client.py "laptop"

# With features and price range
python cli_client.py "laptop" --features "8GB RAM, SSD storage" --min-price 30000 --max-price 80000

# Smartphone search
python cli_client.py "smartphone" --features "good camera, 5G" --min-price 15000 --max-price 50000

Arguments:

• product: Product to search for (required)

• --features, -f: Specific features to look for

• --min-price, -min: Minimum price in rupees

• --max-price, -max: Maximum price in rupees

3. Web Interface (web_client.py)

A beautiful web interface with REST API backend.

# Install additional dependencies
pip install fastapi uvicorn

# Run the web server
python web_client.py

Then open http://localhost:8000 in your browser for a user-friendly interface.

API Endpoints:

• GET /: Web interface

• POST /search: REST API for product search

• GET /health: Health check

4. MCP CLI Integration

You can also use the MCP CLI to interact with your server:

# Install MCP CLI if not already installed
pip install mcp

# Connect to your server
mcp connect stdio -- python main.py

Usage Examples

Example 1: Laptop Search

{
    "product": "laptop",
    "specific_features": "8GB RAM, SSD storage, good for programming",
    "min_price": 30000,
    "max_price": 80000
}

Example 2: Smartphone Search

{
    "product": "smartphone",
    "specific_features": "good camera, long battery life, 5G support",
    "min_price": 15000,
    "max_price": 50000
}

Example 3: Budget Headphones

{
    "product": "wireless headphones",
    "specific_features": "noise cancellation, comfortable",
    "min_price": 1000,
    "max_price": 5000
}

Configuration

API Keys Required

Make sure you have:

1. Hugging Face API Token: Update HF_API_TOKEN in server/buy.py

2. RapidAPI Key: Update the x-rapidapi-key in server/buy.py

Customization

You can customize the AI recommendation prompt in the decision_agent function in server/buy.py.

Integration with Claude Desktop

To use this MCP server with Claude Desktop, add this configuration to your claude_desktop_config.json:

{
  "mcpServers": {
    "amazon-search": {
      "command": "python",
      "args": ["path/to/your/main.py"],
      "env": {}
    }
  }
}

Troubleshooting

Common Issues

1. Import errors: Make sure you're in the correct virtual environment

2. API failures: Check your API keys and internet connection

3. Connection issues: Ensure the MCP server is running before starting clients

Error Messages

• "No result": Usually indicates API issues or no products found

• "Connection refused": MCP server is not running

• "Tool not found": Server initialization issue

Development

Adding New Features

1. Add new tools in server/buy.py using the @mcp.tool() decorator

2. Update client code to use new tools

3. Test with the interactive client first

Testing

# Test the server directly
python -c "from server.buy import mcp; print('Server loads successfully')"

# Test with the interactive client
python client.py

Architecture

┌─────────────────┐    ┌─────────────────┐    ┌─────────────────┐
│   MCP Client    │◄──►│   MCP Server    │◄──►│  External APIs  │
│  (Your Choice)  │    │   (FastMCP)     │    │ (Amazon/HF AI)  │
└─────────────────┘    └─────────────────┘    └─────────────────┘
       │                        │                        │
       ▼                        ▼                        ▼
┌─────────────────┐    ┌─────────────────┐    ┌─────────────────┐
│  • client.py    │    │ • Tool: getdata │    │ • Amazon Search │
│  • cli_client   │    │ • AI Agent      │    │ • HuggingFace   │
│  • web_client   │    │ • FastMCP       │    │ • Recommendations│
│  • Claude       │    │                 │    │                 │
└─────────────────┘    └─────────────────┘    └─────────────────┘

License

This project is open source. Please ensure you comply with the terms of service of the APIs used (Amazon, RapidAPI, Hugging Face).