# MCP Protocol

Anthropic has open-sourced a protocol called [MCP (Model Context Protocol)](https://modelcontextprotocol.io/introduction), which provides a unified and open standard for connecting AI systems with data sources, replacing fragmented integration approaches with a single protocol.

As a result, AI systems can access the data they need in a simpler and more reliable way. With this protocol, developers no longer need to write customized code for different data sources. Instead, they can integrate multiple data sources through standardized MCP interfaces, significantly reducing integration complexity and improving reliability.

This “plug-and-play” capability enables enterprises to quickly inject business data flows into large model training and inference processes, accelerating knowledge acquisition and pattern discovery.

# MCP Architecture Overview

![Architecture Diagram](https://cdn.udelivrs.com/2026/04/3df222261969e945e48456284cf8435d_1776316805502.png)

- **MCP Hosts**: Programs such as Claude Desktop, IDEs, or AI tools that want to access data through MCP  
- **MCP Clients**: Protocol clients that maintain a one-to-one connection with MCP servers  
- **MCP Servers**: Lightweight programs that provide specific capabilities through the standard Model Context Protocol  
- **Local Data Sources**: Files, databases, and services on the local machine that MCP servers can securely access  
- **Remote Services**: External systems on the internet that MCP servers can connect to (e.g., via APIs)