What is Mcp Demonstration

The mcp-demonstration is a demonstration of the Model Context Protocol (MCP) developed by Anthropic. It enables servers to provide resources, tools, and prompts to enhance the capabilities of language models.

Use cases

Use cases for mcp-demonstration include automating responses in customer service, generating code snippets for developers, and assisting users with specific tasks through guided prompts.

How to use

To use mcp-demonstration, install the required dependencies such as NodeJS and Claude Desktop, and configure the settings as specified in the README. You can then access the MCP functionalities through the provided Python SDK or example servers.

Key features

Key features of mcp-demonstration include the ability to provide file-like resources, callable tools for language models with user approval, and pre-written prompts that assist users in completing specific tasks.

Where to use

mcp-demonstration can be used in various fields such as artificial intelligence, software development, and data processing, where enhanced interaction with language models is required.

Model Context Protocol (MCP)

Image Source: https://www.danvega.dev/images/blog/2025/03/11/what_is_mcp.jpeg

Core MCP Concepts

MCP servers can provide three main types of capabilities:
Resources: File-like data that can be read by clients (like API responses or file contents)
Tools: Functions that can be called by the LLM (with user approval)
Prompts: Pre-written templates that help users accomplish specific tasks

MCP: https://modelcontextprotocol.io/introduction
Python SDK: https://github.com/modelcontextprotocol/python-sdk
Example Servers: https://modelcontextprotocol.io/examples
Clients: https://modelcontextprotocol.io/clients
Library: https://pypi.org/project/mcp/

Install

1. NodeJS: https://nodejs.org/en/download
2. Claude Desktop: https://claude.ai/download
3. UV: An extremely fast Python package and project manager, written in Rust. https://docs.astral.sh/uv/ & https://pypi.org/project/uv/
   3.1. pip install uv
   3.2. powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex" for Windows OS
   3.3- curl -LsSf https://astral.sh/uv/install.sh | sh for Linux Base OS

Configuration

Claude Desktop (for Windows)

Path: %APPDATA%/Claude/
Setting Path: %APPDATA%/Claude/claude_desktop_config.json
Logs Path: %APPDATA%/Claude/logs
Please browse to cfg/claude_desktop_config.json file

Cursor (for Windows)

Path: .cursor/
Ctrl + Shift + P
Open MCP Settings
Add new global MCP server
Please browse to cfg/mcp.json file

Windsurf (for Windows)

Path: .codeium/
Ctrl + Shift + P
Open MCP Settings
Add new global MCP server
Please browse to cfg/mcp_config.json file

Examples

https://smithery.ai/
https://smithery.ai/server/@wonderwhy-er/desktop-commander & https://desktopcommander.app/

References

Introducing the Model Context Protocol - https://www.anthropic.com/news/model-context-protocol
Model Context Protocol (MCP) - https://docs.anthropic.com/en/docs/agents-and-tools/mcp
MCP Specification - https://modelcontextprotocol.io/specification/2025-03-26
Microsoft MCP for Beginner - https://github.com/microsoft/mcp-for-beginners
Model Context Protocol (MCP) an overview - https://www.philschmid.de/mcp-introduction
EP154: What is MCP? https://blog.bytebytego.com/p/ep154-what-is-mcp
EP165: AI Agent versus MCP - https://blog.bytebytego.com/p/ep165-ai-agent-versus-mcp
EP163: 12 MCP Servers You Can Use in 2025 - https://blog.bytebytego.com/p/ep163-12-mcp-servers-you-can-use