What is MCP? My Beginner's Guide to Model Context Protocol

Uma Baleboyina — Fri, 08 May 2026 07:51:19 +0000

Introduction

When I started learning Generative AI, one of the first things I came across was something called MCP. At first it sounded complex, but once I understood it, everything clicked. In this blog, I want to share my understanding in the simplest way possible.

What is MCP?
MCP stands for Model Context Protocol. It is a lightweight protocol developed by Anthropic — the company that also created Claude AI. In simple terms, MCP acts as a bridge between an AI assistant and the outside world — connecting it to functions, tools, and external environments so the AI can actually do things, not just talk.

Where did the idea come from?
MCP was inspired by LSP — the Language Server Protocol. If you have used VS Code, you already benefited from LSP without knowing it. LSP is the reason VS Code can support multiple programming languages like Python, JavaScript, Java and more — all through one common standard.
MCP follows the same idea. Instead of building a separate integration for every AI model, MCP provides one common standard that works across multiple models. You build the server once, and any MCP-compatible AI model can use it.

How do you build an MCP server?
Building your first MCP server is simpler than it sounds. You start by importing MCP from the FastMCP library. Then you define your tools using the @mcp.tool decorator and write the logic — the actual code that performs the action. Finally, you connect it to Claude and your server is ready.

The two main components of MCP
MCP has two core components:

MCP Client — This is the intelligent side, like Claude. It understands the user's request and decides what needs to be done.
MCP Server — This is the action side. It actually performs the tasks that the client requests.

Think of the client as the brain and the server as the hands.

How does the workflow work?
Here is the step-by-step flow of how MCP works in action:

Tools vs Resources vs Prompts — what is the difference?
This is one of the most important concepts in MCP:

Tools — These are used to perform actions. When a user wants something done — like fetching weather data or saving a file — a tool does it.
Resources — These are used to feed data to the agent. They are read-only information sources the AI can access when it needs context or data.
Prompts — These are reusable question templates. They represent the general questions or instructions a user commonly asks the AI.

Conclusion
MCP is a powerful and elegant idea — one standard protocol that connects any AI model to any tool or data source. It removes the need to build custom integrations for every model and makes AI agents truly useful in the real world. I am just getting started with MCP, and I am excited to build more servers and share what I learn along the way.

DEV Community: Uma Baleboyina

What is MCP? My Beginner's Guide to Model Context Protocol