AI Glossary
MCP (Model Context Protocol)
Model Context Protocol, MCP protocol, model context protocol
MCP is an open protocol that connects language models to external tools, data and services in a standardized way — through a single interface instead of a separate integration for every application.
- Standardizes how a model connects to tools and data sources.
- Works on a client–server basis: the application calls MCP servers.
- Cuts down the number of separate integrations maintained by hand.
Without a common standard, every connection between a model and a database, file or API requires its own code. MCP defines a single protocol: the application (the client) asks MCP servers for the available tools and data, and the model uses them in a predictable way.
In practice, MCP simplifies building assistants and agents, because adding a new source comes down to connecting another server. Access boundaries and permissions are best set on the server side, before the model starts using it.
Related terms
In guides
Related articles