Model Context Protocol (MCP)
Also known as: MCP
In short
An open protocol that gives AI applications a common interface for accessing external tools, data, and other capabilities.
Model Context Protocol (MCP) standardizes how an AI application connects to external systems. An MCP server can expose capabilities such as retrieving records or calling a business tool. A compatible client makes those capabilities available within the application.
This reduces the need to design a different integration interface for every client and service. It does not mean every client supports every server capability.
A practical example
Imagine an MCP server exposing a read-only product lookup. A specialist asks an assistant which jackets lack material information. The application calls the lookup tool and uses the returned records to answer.
The server does not need to expose price editing for that task. Adding a separate editing tool would require its own implementation and permissions. This is an illustrative integration, not a description of a particular product’s access settings.
What MCP does not provide
MCP is a connection protocol, not a payment method or a promise of visibility in AI shopping results. Commerce protocols can use MCP as an interface while defining their own checkout requirements.
Check the actual tools, authentication, and read or write permissions before connecting a service. Protocol compatibility alone does not establish that an application should access all of your business data.