MCP, or Model Connection Protocol, was introduced by Anthropic to simplify how Language Learning Models (LLMs) connect to external data and tools.
MCP acts as a universal adapter for AI, providing a standard and streamlined way for models to integrate with the broader ecosystem.
By offering a simplified connection process, MCP aims to reduce development and maintenance costs associated with building custom connectors for each data source.
MCP is part of a typical architecture, along with the Agent and RAG, working together like a modern library system to enhance AI capabilities.