What is MCP?
Model Context Protocol (MCP) is an open standard for connecting AI applications to external data sources, tools and workflows through one common interface.
In more detail
Before MCP, every AI app needed its own connector for each system. With MCP, a system exposes a server once, and any compatible client can use it. The project's documentation compares it to a USB-C port for AI applications. Servers can offer tools (actions), resources (data) and prompts. The convenience has a security side. Connecting an agent to a server gives it every capability that server exposes, so review each one, scope its credentials narrowly and treat text coming from a server as untrusted, since it can carry prompt injection. For a single workflow touching two or three systems, a direct API integration is often simpler and easier to audit.
Example
An operations team exposes its ticketing system through one MCP server. Their internal assistant and an engineer's coding tool both use it, with read-only credentials for most people and write access for the on-call lead.