|
|
<!DOCTYPE html> |
|
|
<html lang="en"> |
|
|
<head> |
|
|
<meta charset="UTF-8"> |
|
|
<meta name="viewport" content="width=device-width, initial-scale=1.0"> |
|
|
<title>MCP Architecture - Model Context Protocol</title> |
|
|
<meta name="description" content="Understand the technical architecture of the Model Context Protocol (MCP) including client-server model, components, and implementation details."> |
|
|
<link rel="stylesheet" href="css/styles.css"> |
|
|
<link rel="preconnect" href="https://fonts.googleapis.com"> |
|
|
<link rel="preconnect" href="https://fonts.gstatic.com" crossorigin> |
|
|
<link href="https://fonts.googleapis.com/css2?family=Inter:wght@400;500;600;700&display=swap" rel="stylesheet"> |
|
|
<style> |
|
|
.architecture-diagram { |
|
|
width: 100%; |
|
|
max-width: 800px; |
|
|
margin: 2rem auto; |
|
|
box-shadow: 0 4px 6px rgba(0, 0, 0, 0.1); |
|
|
border-radius: 8px; |
|
|
} |
|
|
</style> |
|
|
</head> |
|
|
<body> |
|
|
<header> |
|
|
<div class="container header-container"> |
|
|
<a href="index.html" class="logo">MCP<span>Hub</span></a> |
|
|
<nav> |
|
|
<ul> |
|
|
<li><a href="index.html">Home</a></li> |
|
|
<li><a href="about.html">About</a></li> |
|
|
<li><a href="architecture.html">Architecture</a></li> |
|
|
<li><a href="benefits.html">Benefits</a></li> |
|
|
<li><a href="ecosystem.html">Ecosystem</a></li> |
|
|
<li><a href="getting-started.html">Get Started</a></li> |
|
|
<li><a href="faq.html">FAQ</a></li> |
|
|
</ul> |
|
|
</nav> |
|
|
</div> |
|
|
</header> |
|
|
|
|
|
<section class="hero"> |
|
|
<div class="container"> |
|
|
<h1>MCP Architecture</h1> |
|
|
<p>Understanding the technical structure and components of the Model Context Protocol</p> |
|
|
</div> |
|
|
</section> |
|
|
|
|
|
<main class="container"> |
|
|
<section> |
|
|
<h2>Client-Server Architecture</h2> |
|
|
<p> |
|
|
The Model Context Protocol (MCP) follows a client-server architecture that enables AI models to connect with external data sources and tools. This architecture consists of two main components: |
|
|
</p> |
|
|
|
|
|
|
|
|
<div class="architecture-diagram"> |
|
|
<img src="images/mcp_architecture_diagram.svg" alt="MCP Architecture Diagram showing clients, servers, and transport layer interactions" width="800" height="600"> |
|
|
</div> |
|
|
|
|
|
<div class="two-column"> |
|
|
<div> |
|
|
<h3>MCP Clients</h3> |
|
|
<p> |
|
|
MCP clients are applications that use AI models and need to access external data or functionality. Examples include: |
|
|
</p> |
|
|
<ul style="margin-left: 2rem; margin-bottom: 1.5rem;"> |
|
|
<li>AI-powered chat interfaces (like Claude Desktop)</li> |
|
|
<li>Integrated development environments (IDEs)</li> |
|
|
<li>Custom applications that incorporate LLMs</li> |
|
|
</ul> |
|
|
<p> |
|
|
Clients connect to MCP servers to retrieve context, execute tools, and access data needed by the AI model. |
|
|
</p> |
|
|
</div> |
|
|
<div> |
|
|
<h3>MCP Servers</h3> |
|
|
<p> |
|
|
MCP servers expose data sources, tools, and functionality to MCP clients. Examples include: |
|
|
</p> |
|
|
<ul style="margin-left: 2rem; margin-bottom: 1.5rem;"> |
|
|
<li>File system servers (for accessing local files)</li> |
|
|
<li>Database connectors (for querying databases)</li> |
|
|
<li>API integrations (for accessing web services)</li> |
|
|
<li>Custom tool implementations (for specialized functionality)</li> |
|
|
</ul> |
|
|
<p> |
|
|
Servers handle requests from clients, process them, and return results back to the client. |
|
|
</p> |
|
|
</div> |
|
|
</div> |
|
|
</section> |
|
|
|
|
|
<section> |
|
|
<h2>Communication Flow</h2> |
|
|
<p> |
|
|
The communication between MCP clients and servers follows a structured flow: |
|
|
</p> |
|
|
<ol style="margin-left: 2rem; margin-bottom: 1.5rem;"> |
|
|
<li> |
|
|
<strong>Connection Establishment:</strong> The client initiates a connection to one or more MCP servers. |
|
|
</li> |
|
|
<li> |
|
|
<strong>Tool Discovery:</strong> The client queries the server for available tools and capabilities using the <code>listTools</code> method. |
|
|
</li> |
|
|
<li> |
|
|
<strong>Tool Invocation:</strong> The client requests the server to perform an action using the <code>callTool</code> method, passing the necessary parameters. |
|
|
</li> |
|
|
<li> |
|
|
<strong>Response Processing:</strong> The server executes the requested action and returns the results to the client. |
|
|
</li> |
|
|
<li> |
|
|
<strong>Model Integration:</strong> The client incorporates the server's response into the AI model's context, allowing the model to use the information in generating its response. |
|
|
</li> |
|
|
</ol> |
|
|
<p> |
|
|
This structured communication flow ensures that AI models can access external data and functionality in a consistent and secure manner. |
|
|
</p> |
|
|
</section> |
|
|
|
|
|
<section> |
|
|
<h2>MCP Technical Components</h2> |
|
|
<div class="card-container"> |
|
|
<div class="card"> |
|
|
<h3>Protocol Layer</h3> |
|
|
<p> |
|
|
The protocol layer handles message framing, request/response linking, and high-level communication patterns between clients and servers. It defines the structure of messages and the sequence of operations. |
|
|
</p> |
|
|
</div> |
|
|
<div class="card"> |
|
|
<h3>Transport Layer</h3> |
|
|
<p> |
|
|
The transport layer deals with the actual communication between clients and servers. MCP supports multiple transport mechanisms, including stdio for local processes and HTTP over SSE (Server-Sent Events) for remote connections. |
|
|
</p> |
|
|
</div> |
|
|
<div class="card"> |
|
|
<h3>Tool Specifications</h3> |
|
|
<p> |
|
|
Tool specifications define the capabilities that MCP servers expose to clients. Each tool includes a name, description, parameter schema, and return type, allowing clients to understand how to use the tool. |
|
|
</p> |
|
|
</div> |
|
|
<div class="card"> |
|
|
<h3>Security Model</h3> |
|
|
<p> |
|
|
The security model ensures that MCP connections are secure and that data is protected. It includes features like server-controlled access to resources and clear system boundaries. |
|
|
</p> |
|
|
</div> |
|
|
</div> |
|
|
</section> |
|
|
|
|
|
<section> |
|
|
<h2>Transport Mechanisms</h2> |
|
|
<p> |
|
|
MCP supports different transport mechanisms for communication between clients and servers: |
|
|
</p> |
|
|
<div class="two-column"> |
|
|
<div> |
|
|
<h3>Stdio Servers</h3> |
|
|
<p> |
|
|
Stdio servers run as subprocesses of the client application, enabling local communication through standard input/output streams. This approach is suitable for accessing local resources like the file system or for running tools directly on the client machine. |
|
|
</p> |
|
|
<p> |
|
|
Stdio servers are easy to set up and secure for local operations but are limited to running on the same machine as the client. |
|
|
</p> |
|
|
</div> |
|
|
<div> |
|
|
<h3>HTTP over SSE Servers</h3> |
|
|
<p> |
|
|
HTTP over Server-Sent Events (SSE) enables remote communication between clients and servers. This approach allows MCP servers to run on different machines or in cloud environments, making it suitable for distributed architectures. |
|
|
</p> |
|
|
<p> |
|
|
HTTP/SSE servers provide greater flexibility for deployment but require additional security considerations for remote access. |
|
|
</p> |
|
|
</div> |
|
|
</div> |
|
|
</section> |
|
|
|
|
|
<section> |
|
|
<h2>Tool Implementation</h2> |
|
|
<p> |
|
|
MCP tools are the functional units that MCP servers expose to clients. Each tool has specific characteristics: |
|
|
</p> |
|
|
<ul style="margin-left: 2rem; margin-bottom: 1.5rem;"> |
|
|
<li> |
|
|
<strong>Name:</strong> A unique identifier for the tool. |
|
|
</li> |
|
|
<li> |
|
|
<strong>Description:</strong> A human-readable explanation of what the tool does. |
|
|
</li> |
|
|
<li> |
|
|
<strong>Parameters:</strong> A schema defining the inputs the tool expects, usually in JSON Schema format. |
|
|
</li> |
|
|
<li> |
|
|
<strong>Return Value:</strong> The type and structure of data the tool returns. |
|
|
</li> |
|
|
</ul> |
|
|
<p> |
|
|
Tools can range from simple utilities like file readers to complex integrations with external APIs or custom business logic. |
|
|
</p> |
|
|
<pre><code>// Example tool definition |
|
|
{ |
|
|
"name": "readFile", |
|
|
"description": "Reads the contents of a file", |
|
|
"parameters": { |
|
|
"type": "object", |
|
|
"properties": { |
|
|
"path": { |
|
|
"type": "string", |
|
|
"description": "Path to the file to read" |
|
|
} |
|
|
}, |
|
|
"required": ["path"] |
|
|
}, |
|
|
"returnType": { |
|
|
"type": "string", |
|
|
"description": "Contents of the file" |
|
|
} |
|
|
}</code></pre> |
|
|
</section> |
|
|
|
|
|
<section> |
|
|
<h2>Implementation Libraries</h2> |
|
|
<p> |
|
|
Anthropic and the MCP community provide libraries to simplify the implementation of MCP clients and servers: |
|
|
</p> |
|
|
<div class="two-column"> |
|
|
<div> |
|
|
<h3>Client Libraries</h3> |
|
|
<p> |
|
|
Client libraries handle the complexity of connecting to MCP servers, discovering tools, and making tool calls. They are available for different programming languages and environments. |
|
|
</p> |
|
|
<ul style="margin-left: 2rem; margin-bottom: 1.5rem;"> |
|
|
<li>TypeScript/JavaScript MCP Client</li> |
|
|
<li>Python MCP Client</li> |
|
|
<li>Java MCP Client (in collaboration with Spring AI)</li> |
|
|
<li>C# MCP Client</li> |
|
|
</ul> |
|
|
</div> |
|
|
<div> |
|
|
<h3>Server Libraries</h3> |
|
|
<p> |
|
|
Server libraries simplify the creation of MCP servers by handling protocol implementation, tool registration, and request processing. They allow developers to focus on implementing the tool functionality. |
|
|
</p> |
|
|
<ul style="margin-left: 2rem; margin-bottom: 1.5rem;"> |
|
|
<li>TypeScript/JavaScript MCP Server</li> |
|
|
<li>Python MCP Server</li> |
|
|
<li>Java MCP Server</li> |
|
|
<li>C# MCP Server</li> |
|
|
</ul> |
|
|
</div> |
|
|
</div> |
|
|
</section> |
|
|
|
|
|
<section> |
|
|
<h2>Security Architecture</h2> |
|
|
<p> |
|
|
Security is a fundamental aspect of the MCP architecture, with several key principles: |
|
|
</p> |
|
|
<ul style="margin-left: 2rem; margin-bottom: 1.5rem;"> |
|
|
<li> |
|
|
<strong>Server-Controlled Access:</strong> MCP servers maintain control over their resources, determining what data is accessible to clients. |
|
|
</li> |
|
|
<li> |
|
|
<strong>No API Key Sharing:</strong> There's no need to share API keys with LLM providers, as the MCP server handles authentication with external services. |
|
|
</li> |
|
|
<li> |
|
|
<strong>Clear System Boundaries:</strong> The client-server architecture establishes clear boundaries between systems, limiting exposure of sensitive information. |
|
|
</li> |
|
|
<li> |
|
|
<strong>Local Processing:</strong> For sensitive operations, MCP can use local stdio servers to ensure data doesn't leave the user's machine. |
|
|
</li> |
|
|
</ul> |
|
|
<p> |
|
|
As MCP evolves, additional security features are being developed, including enterprise-grade authentication for remote servers and more fine-grained access controls. |
|
|
</p> |
|
|
</section> |
|
|
</main> |
|
|
|
|
|
<footer> |
|
|
<div class="container footer-container"> |
|
|
<div> |
|
|
<h4>MCP Resources</h4> |
|
|
<ul> |
|
|
<li><a href="https://modelcontextprotocol.io" target="_blank">Official Documentation</a></li> |
|
|
<li><a href="https://github.com/modelcontextprotocol" target="_blank">GitHub Repository</a></li> |
|
|
<li><a href="https://www.anthropic.com/news/model-context-protocol" target="_blank">Anthropic MCP Announcement</a></li> |
|
|
</ul> |
|
|
</div> |
|
|
<div> |
|
|
<h4>Learn More</h4> |
|
|
<ul> |
|
|
<li><a href="about.html">About MCP</a></li> |
|
|
<li><a href="architecture.html">Architecture</a></li> |
|
|
<li><a href="benefits.html">Benefits</a></li> |
|
|
<li><a href="ecosystem.html">Ecosystem</a></li> |
|
|
</ul> |
|
|
</div> |
|
|
<div> |
|
|
<h4>Community</h4> |
|
|
<ul> |
|
|
<li><a href="https://github.com/modelcontextprotocol/discussions" target="_blank">Discussions</a></li> |
|
|
<li><a href="https://github.com/modelcontextprotocol/community-servers" target="_blank">Community Servers</a></li> |
|
|
</ul> |
|
|
</div> |
|
|
</div> |
|
|
<div class="copyright container"> |
|
|
<p>© 2025 MCP Information Hub. Model Context Protocol is an open source project developed by Anthropic, PBC.</p> |
|
|
</div> |
|
|
</footer> |
|
|
|
|
|
<script src="js/main.js"></script> |
|
|
</body> |
|
|
</html> |
