Anthropic MCP

protocol
https://modelcontextprotocol.io

The Model Context Protocol (MCP) by Anthropic is an open standard that defines how AI applications connect to external tools and data. It provides a client-server architecture where MCP clients (AI apps) discover and use tools exposed by MCP servers.

MCP standardizes tool discovery, invocation, and response handling. A tool written as an MCP server works with any MCP-compatible client — Claude Desktop, VS Code extensions, agent frameworks, and more. This eliminates the need for custom integrations per tool per client.

The protocol also supports resources (read-only data sources) and prompts (reusable prompt templates), making it a comprehensive interface between AI and external systems.

Key Features

  • Standardized tool discovery and invocation
  • Client-server architecture
  • Resources for read-only data access
  • Prompt templates
  • Growing ecosystem of MCP servers

Pros

  • +Open standard — no vendor lock-in
  • +Write once, use everywhere
  • +Growing ecosystem
  • +Backed by Anthropic

Cons

  • Still relatively new
  • Not all tools have MCP servers yet
  • Some frameworks have limited MCP support