AI agent glossary
MCP client
The part of an AI application that connects to MCP servers, discovers their tools and resources, and passes them to the model.
In the Model Context Protocol, the client lives inside the host application, such as Claude Code, the Claude desktop app, an IDE, or your own agent. It opens a connection to each configured MCP server, asks what tools, resources, and prompts the server offers, and makes them available to the model.
When the model decides to call one of those tools, the client sends the request to the right server and returns the result to the model. The host typically handles permissions and approval prompts at this layer.
If you are using agents, you usually configure clients rather than build them. If you are building your own agent, an agent SDK or the official MCP SDKs can provide the client for you.
Learn more in the course
Related terms
Go from definitions to doing
Learn to use and build AI agents in 65 practical lessons.