MCP (Model Context Protocol)

A standard that enables AI models to interact with external tools and context seamlessly.

How it works

The Model Context Protocol establishes a standardised way for an AI agent to discover and invoke external tools. Instead of hardcoding API integrations for every possible tool, a developer builds an MCP server that exposes tools with a standard schema. The AI model, acting as an MCP client, can then dynamically list the available tools and call them by emitting structured function calls. MCP servers can expose file systems, databases, web search, code execution, and any other capability.

Why it matters

MCP solves the fragmentation problem in AI tool use. Before MCP, every AI assistant required custom integrations for every tool — a rapidly unmanageable combinatorial problem as the number of models and tools grew. By standardising the communication layer, MCP allows developers to build a tool once and have it work with any compliant model. This dramatically accelerates the development of capable, autonomous AI agents.