MCP glossary
The Model Context Protocol vocabulary, defined in plain English — each term in one quotable sentence, then why it actually matters. If you're new to MCP, start with MCP itself and MCP server.
- Model Context Protocol (MCP)
- The Model Context Protocol (MCP) is an open standard that lets AI applications connect to external tools and data through a single, uniform interface.
- MCP server
- An MCP server is a program that exposes a set of tools (and optionally resources and prompts) to an AI client over the Model Context Protocol.
- MCP client
- An MCP client is the component inside an AI application that connects to MCP servers and makes their tools available to the model.
- MCP host
- The MCP host is the AI application (e.g. Claude Desktop or Cursor) that contains the model and the MCP client, and decides when to call tools.
- stdio transport
- stdio transport runs an MCP server as a local process and communicates with it over standard input/output — no network involved.
- HTTP / SSE transport
- HTTP (streamable) and SSE transports run an MCP server as a hosted endpoint you connect to over the network via a URL.
- MCP tool
- An MCP tool is a single callable function an MCP server exposes — with a name, a description, and a typed input schema the model uses to call it.
- MCP resource
- An MCP resource is read-only data an MCP server exposes for the model to reference — like a file, a record, or a document — addressed by a URI.
- MCP prompt
- An MCP prompt is a reusable, parameterized message template an MCP server offers, which a user can invoke to start a task with the model.
- mcp-remote (proxy)
- mcp-remote is a small proxy that lets clients without native remote support (notably Claude Desktop) connect to a remote HTTP/SSE MCP server.
- tool budget
- The tool budget is the practical limit on how many active MCP tools a client can handle before tool selection starts to degrade — roughly 40 in Cursor.
- JSON-RPC (in MCP)
- MCP is built on JSON-RPC 2.0, a lightweight remote-procedure-call protocol that encodes requests and responses as JSON messages.