MCP Server
Also called Model Context Protocol server, MCP connector
A program that offers tools, data or prompts to AI applications through the Model Context Protocol, so any compatible app can use them.
What is an MCP server?
An MCP server is a small program that connects an AI application to something else, such as a database, a code repository, a calendar or a company's API, using the Model Context Protocol (MCP). Once an AI app such as a chat assistant or coding tool is connected to it, the AI can use what the server offers.
Because MCP is a shared standard, one server works with any app that supports MCP, instead of each tool needing a custom integration for each app.
How does it work?
The AI application (the host) runs an MCP client that connects to each server. A server can offer three kinds of thing:
- Tools: actions the AI can ask to run, such as "create an issue" or "query the sales table", through tool calling.
- Resources: data the app can read, such as files or records.
- Prompts: ready-made instructions or templates for common tasks.
Servers can run on your own computer, talking to the app directly, or remotely over the internet, usually with a sign-in.
What is it commonly confused with?
- The protocol itself. MCP is the standard; an MCP server is one implementation of it for a particular service.
- Agent skills. Skills package instructions and know-how; MCP servers provide connections to tools and data. Many setups use both.
- A normal API. Most MCP servers wrap an existing API and describe it in a way AI apps can discover and use.
Why does it matter?
MCP servers are how AI assistants get access to your real tools and data. They also carry real permissions, so only install servers from sources you trust, grant the least access they need, and be aware that content a server returns can contain prompt injection. For a plain-English overview, read MCP 101.
