This repository contains a collection of example Model Context Protocol (MCP) servers, each demonstrating different integrations and use cases for the MCP standard. These servers are designed for educational purposes as part of the course "MCP for beginners", on KodeKloud and can be used as templates or references for building your own MCP-compatible tools.
Purpose:
A full-featured MCP server that integrates with Google Calendar, allowing AI assistants (like Claude) to manage calendar events, check availability, and perform smart scheduling.
Key Features:
- Multi-calendar support
- Create, update, delete, and search events
- Recurring event management
- Free/busy queries across calendars
- Smart scheduling and natural language date parsing
- Import events from images, PDFs, or web links
How it works:
The server uses Google OAuth for authentication and communicates with the Google Calendar API. It exposes a set of MCP tools for calendar operations, which can be accessed by MCP clients.
Setup:
- Requires a Google Cloud project and OAuth credentials
- Can be run via
npx, locally, or with Docker - See the google-calendar-mcp/README.md and docs/ for detailed setup, usage, and deployment instructions
Purpose:
A minimal MCP server example that implements a single tool: integer addition.
Key Features:
- One tool:
add-integers(adds two integers) - Demonstrates basic STDIO transport and JSON-RPC handling
- Includes comprehensive tests for input validation and error handling
How it works:
The server listens for MCP requests over STDIO, validates input, performs integer addition, and returns the result.
Setup:
- Run with
node server.js - Test with
npm test - See node-basic-mcp-server/README.md for details
Purpose:
An MCP server that connects to a local Ollama instance, exposing structured prompt-based tools (like code review, concept explanation, and greetings) to MCP clients.
Key Features:
- Integrates with Ollama's local LLM API
- Provides prompt-based tools: greeting, code review, concept explanation
- Uses STDIO transport for MCP communication
How it works:
The server registers several prompts as MCP tools. When a tool is called, it renders a prompt template, sends it to Ollama, and returns the LLM's response.
Setup:
- Requires Ollama running locally (
ollama serve) - Start with
node server.js - Prompts are defined in the
prompts/directory - See code and comments in ollama-mcp/server.js for details
Purpose:
A demonstration MCP server for Ollama, similar to ollama-mcp, but with additional demo scripts and configuration examples for integration with Claude Desktop.
Key Features:
- Same prompt-based tools as
ollama-mcp(greeting, code review, concept explanation) - Includes demo/test scripts to showcase prompt execution
- Provides example configuration files for Claude Desktop integration
How it works:
The server exposes prompts as MCP tools, renders templates with user arguments, sends them to Ollama, and returns the results. Demo scripts (demo.js, comprehensive-demo.js, etc.) show how to interact with the server programmatically.
Setup:
- Requires Ollama running locally
- Start with
node server.js - See USAGE.md and SETUP.md for integration instructions with Claude Desktop
Purpose:
Auto-generated MCP servers using the Postman MCP Generator, exposing API requests as MCP tools.
Key Features:
- Each subdirectory (e.g.,
postman-mcp-server,postman-mcp-server 2) is a separate generated server - Tools are generated from selected Postman API requests
- Supports environment variable configuration for API keys
- Docker support for deployment
How it works:
Each server exposes API endpoints as MCP tools, allowing MCP clients to call external APIs via a standardized interface.
Setup:
- Install dependencies with
npm install - Run with
node mcpServer.js - Configure environment variables as needed
- See the respective README.md for full setup and usage
Purpose:
An MCP server that provides weather data using the Open-Meteo API, allowing AI assistants to retrieve current weather and forecasts.
Key Features:
- Get current weather, hourly forecasts, and weather summaries
- Global coverage (any latitude/longitude)
- No API key required (uses free Open-Meteo API)
- Comprehensive error handling
How it works:
The server exposes weather-related tools via MCP, fetching data from Open-Meteo and returning it in a structured format.
Setup:
- Install dependencies with
npm install - Start the server with
npm start - See Weather-MCP-Server/README.md for details
All code is provided under the MIT License unless otherwise specified in subdirectories.
Contributions and suggestions are welcome! Please open issues or pull requests for improvements or new examples.
If you need help with a specific example, refer to the README or documentation within each subdirectory.