MCP server
Definition
An MCP server is a small program that connects an AI assistant to one tool or data source, such as Notion, Gmail or a database, using the Model Context Protocol. It tells the assistant what it can read or do there and carries out those actions when asked.
By Ronan Liedmeier, fractional head of growth and course author · Published 17 October 2026 · Last updated 17 October 2026
Also called: MCP connector
Each server you connect lets the assistant into a real account with your data in it.
How it works
In MCP terms, the AI app you use (Claude Desktop, Claude Code or VS Code, for example) is the host, and it opens a separate connection to each server. The server lists what it offers: tools the AI can call, such as "create page", resources it can read, such as a file or a database schema, and reusable prompt templates. A local server runs on your own computer. A remote server runs on the vendor's side and you sign in to it, and the MCP docs recommend OAuth for that sign-in.
Example
A marketing manager connects her CRM's official MCP server to her assistant with read access only. She asks for every open deal that has had no logged activity for 30 days. For two weeks she checks the list against the CRM. Once it matches, she allows one write action: adding a follow-up task to a deal, and only after she approves the draft.
Common mistake
Connecting five tools on the first day and accepting whatever scopes each one requests. Unofficial servers often ask for full write access when you only need to read. Connect one tool you use daily, stay read-only until its answers check out, and disconnect servers you no longer use.
Learn it in AI Academy
- Connecting One MCP Server (Consumer)Advanced · 14 min
- MCP: The Integration LayerExpert · 8 min
Related terms
- Model Context Protocol (MCP)
- Model Context Protocol (MCP) is an open standard that lets AI apps connect to outside tools and data, such as a calendar, a CRM or a database, through one shared format. Apps that support MCP can then use those tools without a custom integration for each one.
- MCP gateway
- An MCP gateway is a single entry point that sits between your AI agents and all of your MCP servers. It holds the passwords and keys, decides which tools each agent may use, and keeps one log of every call.
- Tool use (function calling)
- Tool use, also called function calling, is when an AI model asks your software to run a specific action, such as a search or a database lookup, and then uses the result in its answer. The model chooses the tool and the inputs; your code runs it.
- API key
- An API key is a long secret code that tells a software service which account is making a request. Anyone who has the key can act as you, within what that key is allowed to do, so it needs the same care as a password.