What is MCP?
Model Context Protocol (MCP) is an open standard developed by Anthropic in 2024, enabling AI models to securely connect with external data sources and tools. It's often called "USB-C for AI" — one universal interface instead of dozens of custom integrations.
How does MCP work?
MCP defines three resource types a server can expose to an AI agent: tools (e.g., database search, email sending), resources (files, documents, data), and prompts (instruction templates). The agent connects to an MCP server and dynamically discovers available capabilities.
Why is MCP transformative?
Before MCP, every AI integration required custom code. MCP standardizes this: one MCP server for Slack works with any agent supporting the protocol — Claude, GPT, local models. By 2026, over 10,000 MCP servers exist, and all major AI providers support the protocol.