AI Infrastructure

What Is an MCP Server?

A practical explanation of MCP servers, clients, tools, resources, prompts, authentication, and security.

An MCP server is a program that exposes tools, resources, and prompts to MCP clients. It is the bridge between an AI assistant and a real system such as GitHub, Slack, a database, a filesystem, a browser, or an internal app.

The server does not replace the model. It gives the model a controlled way to see context or request actions.

Quick Verdict

MCP server roleExpose capabilities

Common capabilitiesTools, resources, prompts

Can wrapAPIs, files, databases, apps, workflows

Security focusLeast privilege and human approval

What an MCP Server Provides

A server can provide tools for actions, resources for readable context, and prompts for reusable task patterns. It also declares capabilities so the client knows what the server supports.

Local vs Remote MCP Servers

Local servers often run on a developer's machine and connect through stdio. Remote servers can run behind HTTP-based transports and usually need stronger authentication, authorization, logging, and deployment practices.

Security Basics

An MCP server should expose the smallest useful set of capabilities. Avoid broad filesystem access, avoid unaudited write actions, validate inputs, protect credentials, and log important tool calls.

Direct Answers