Model Context Protocol (MCP) Integration Guide

This guide covers integrating the Prompt Alchemy MCP server with AI assistants and IDEs.

Overview

Prompt Alchemy’s MCP server exposes prompt management capabilities via the Model Context Protocol. It runs as a Docker container providing 19+ tools for prompt generation, search, optimization, and more.

Setup

The server can be run locally or deployed using Docker. See the Deployment Guide for detailed instructions.

Available Tools

The MCP server exposes all 15 core tools of Prompt Alchemy. For a complete list, including request/response schemas, please see the MCP API Reference.

Integrating with Hosts

Claude Desktop

Configure Claude Desktop to run:

docker exec -i prompt-alchemy-mcp prompt-alchemy serve

VS Code / Cursor / Zed

Configure your IDE’s MCP settings to execute the docker exec command above.

Testing Integration

Use the test script:

python3 test-mcp.py

This verifies:

Protocol initialization
Tool listing
Tool calling

Best Practices

Use docker-compose for management
Monitor logs for issues
Update config via mounted docker-config.yaml

For custom integrations, use the JSON-RPC protocol directly as shown in the test script.

MCP Server Integration