eBay MCP Server

eBay MCP Server

A Model Context Protocol (MCP) server for interacting with eBay's APIs. Enables AI assistants to search eBay listings, track prices, and find deals on homelab equipment and other items.

Features

✅ Implemented (8 tools)

• 🔍 Search eBay listings - Basic and advanced search with all filters
• 💰 Price tracking - SQLite database tracks prices over time
• 📊 Deal detection - Analyzes market data to find bargains (15%+ below average)
• 📦 Item details - Comprehensive info including seller reputation
• 📈 Price history - View trends with statistics (min/max/avg/median)
• 📋 Watchlist - Track multiple items with custom alerts

🚧 Planned (12 tools)

• 🏷️ Category browsing - Navigate eBay's category hierarchy
• 🆕 New listings - Fresh deals (last 6-24 hours)
• ⏰ Ending soon - Auction sniper helper
• 📊 Market value - Price research tool

Status

🟢 MVP Complete - 8 core tools implemented and ready for testing!

See IMPLEMENTATION_STATUS.md for detailed progress.

Planned Tools

• search_ebay - Search listings with filters (keywords, price, condition, category)
• get_item_details - Get detailed information about a specific listing
• track_price - Add an item to price tracking
• get_price_history - View historical price data for tracked items
• find_deals - Search for items below market value based on historical data

Requirements

• Python 3.10+
• eBay Developer Account (for API credentials)
• MCP-compatible client (Claude Desktop, Cline, etc.)

Quick Start

1. Get eBay API Credentials

   • Sign up at https://developer.ebay.com/
   • Create an application
   • Copy your App ID

2. Install

   git clone https://github.com/hanku4u/ebay-mcp-server.git
   cd ebay-mcp-server
   pip install -e .
   

3. Configure

   cp .env.example .env
   # Edit .env and add your EBAY_APP_ID
   

4. Test

   python -m ebay_mcp
   # Server will start in stdio mode (MCP standard)
   

Configuration

Set your eBay API credentials as environment variables:

export EBAY_APP_ID="your-app-id"
export EBAY_CERT_ID="your-cert-id"
export EBAY_DEV_ID="your-dev-id"

Or create a .env file:

EBAY_APP_ID=your-app-id
EBAY_CERT_ID=your-cert-id
EBAY_DEV_ID=your-dev-id

Usage

With Claude Desktop

Add to your Claude Desktop config (~/Library/Application Support/Claude/claude_desktop_config.json on macOS):

{
  "mcpServers": {
    "ebay": {
      "command": "python",
      "args": ["-m", "ebay_mcp"],
      "env": {
        "EBAY_APP_ID": "your-app-id",
        "EBAY_CERT_ID": "your-cert-id",
        "EBAY_DEV_ID": "your-dev-id"
      }
    }
  }
}

Alternatively, use the FastMCP CLI:

```json
{
  "mcpServers": {
    "ebay": {
      "command": "fastmcp",
      "args": ["run", "path/to/ebay_mcp/server.py:mcp"],
      "env": {
        "EBAY_APP_ID": "your-app-id",
        "EBAY_CERT_ID": "your-cert-id",
        "EBAY_DEV_ID": "your-dev-id"
      }
    }
  }
}

With OpenClaw

Coming soon - will integrate with OpenClaw's MCP client capabilities.

HTTP Server (Remote Access)

Run as an HTTP server for remote access:

python -m ebay_mcp
# Or with FastMCP CLI:
fastmcp run src/ebay_mcp/server.py:mcp --transport http --port 8000

Then connect from any MCP client:

from fastmcp import Client

async with Client("http://localhost:8000/mcp") as client:
    result = await client.call_tool("search_ebay", {
        "keywords": "Dell PowerEdge R720",
        "max_price": 500
    })
    print(result)

eBay API Access

To use this server, you'll need eBay API credentials:

1. Sign up for the eBay Developers Program
2. Create a new application
3. Generate your App ID, Cert ID, and Dev ID
4. Choose the appropriate API: Finding API (for search) or Trading API (for account management)

Recurring Searches with OpenClaw

Example cron job for daily homelab equipment searches:

{
  "schedule": { "kind": "cron", "expr": "0 9 * * *", "tz": "America/Chicago" },
  "payload": {
    "kind": "agentTurn",
    "message": "Search eBay for homelab servers under $500 and summarize the best deals"
  },
  "sessionTarget": "isolated"
}

Architecture

Built using FastMCP - the fast, Pythonic way to build MCP servers. FastMCP powers 70% of MCP servers and makes it easy to create production-ready integrations with minimal boilerplate.

Example: Adding a Tool

Adding new tools is as simple as decorating a function:

from fastmcp import FastMCP

mcp = FastMCP("eBay MCP Server")

@mcp.tool
def search_ebay(
    keywords: str,
    max_price: float = None,
    condition: str = "New"
) -> dict:
    """Search eBay listings with filters"""
    # Implementation here
    return {...}

if __name__ == "__main__":
    mcp.run()  # Supports stdio and HTTP transports