Wraps the Kapruka.com REST API into MCP tools so Claude can browse and interact with Sri Lanka's largest e-commerce catalog. The server exposes product search, order management, and other shop operations through a FastMCP-based Python implementation. You can run it locally via stdio for desktop clients or deploy the streamable HTTP transport at scale. Useful if you're building shopping assistants, inventory tools, or automated ordering workflows that need programmatic access to Kapruka's marketplace. Includes a developer CLI for health checks and tool inspection, plus full pytest coverage.
Python MCP server that wraps the Kapruka.com REST API and exposes it as tools for LLMs and third-party MCP clients.
# 1. Create and activate a virtual environment
python -m venv .venv
source .venv/bin/activate # Windows: .venv\Scripts\activate
# 2. Install dependencies
pip install -e ".[dev]"
# 3. Configure environment
cp .env.example .env
# Edit .env with your Kapruka API URL and key
# Start the MCP server (streamable HTTP, default port 8000)
python cli.py server
# Start with stdio transport (for use with MCP Inspector)
python cli.py server --stdio
# Health-check the Kapruka REST API
python cli.py ping
# List registered MCP tools
python cli.py tools
npx @modelcontextprotocol/inspector python cli.py server --stdio
src/
server.py # FastMCP server entry point
tools/ # One module per tool group (products, orders, …)
api/
client.py # Async httpx client + error handling
config/
settings.py # Env-based configuration
tests/ # pytest test suite
cli.py # Developer CLI
pytest
io.github.shelvick/shopify-subscription-reconciliation
zleventer/google-ads-mcp
csoai-org/meok-stripe-acp-checkout-mcp
io.github.mharnett/google-ads
csoai-org/stripe-billing-mcp
co.pipeboard/google-ads-mcp