Standardizes client-server communication, allowing large language models to access server-defined tools and resources.
The Model Context Protocol (MCP) establishes a standardized client-server architecture for artificial intelligence models, particularly Large Language Models (LLMs), to effectively communicate with and utilize server-side assets. It provides a structured framework for defining and exposing tools, resources, prompts, and sampling mechanisms, allowing LLMs to extend their capabilities by interacting with external functionalities and data sources in a controlled environment.