ez

EZ-MCP provides battle-tested templates for building MCP servers in Python (uv) or TypeScript (Deno). Each template demonstrates core MCP features such as Resources for dynamic context, Tools that LLMs can call to perform actions, Prompts for reusable interaction patterns, and environment-variable configuration. Start a server with either the Python-based template or the TypeScript/Deno template, then connect MCP clients via an mcp.json configuration. Use the included MCP Inspector during development to browse available tools, test inputs, and view resource contents. You can add your own tools by decorating functions in Python or registering tools in TypeScript, and you can customize server metadata like name and version directly in the template code. The templates also show how to wire up external dependencies, databases, and API integrations to enrich LLM context and tool functionality.

Prerequisites:

• Python 3.9+ with uv installed (or curl to install UV as shown)
• Node.js environment is not required for the Python template, but Deno is used for the TypeScript template

Installation steps (choose Python or TypeScript):

Python template:

1. Install UV (if not already installed): curl -LsSf https://astral.sh/uv/install.sh | sh
2. Clone the repository and navigate to ez-mcp: git clone <repo-url> cd ez-mcp
3. Run the Python MCP server: uv run ez-mcp.py

TypeScript template:

1. Install Deno (if not already installed): curl -fsSL https://deno.land/install.sh | sh
2. Clone the repository and navigate to ez-mcp: git clone <repo-url> cd ez-mcp
3. Run the TypeScript MCP server: deno run --allow-all ez-mcp.ts

Notes:

• Ensure you copy or customize ez-mcp.py / ez-mcp.ts to your environment before running.
• If you plan to connect via an mcp.json client configuration, adjust the command and path to your server accordingly.
• You can install the MCP Inspector to test interactively during development.

Tips and common considerations:

• Environment variables: Use env in your mcp.json to pass API keys or configuration; never hard-code secrets in code.
• Dependencies: In Python, add dependencies like mcp>=1.9.0 or requests; in TypeScript, install needed modules via Deno imports as shown in the templates.
• Tools and resources: Explore the provided examples for file search, web scraping, database queries, and API integrations to understand how to expose functionality to LLMs.
• Testing: Use the MCP Inspector during development to verify tool outputs, resource contents, and prompt flows before deploying.
• Portability: The Python uv template and the TS/Den template are functionally identical; choose the one you are most comfortable with and customize as needed.