What is MCP
MCP (Model Context Protocol) is an open standard that defines how AI applications connect to external tools, data sources, and workflows. Instead of building custom integrations for every tool, developers implement one protocol and get access to any MCP-compatible service.
How it works
MCP follows a client-server architecture. An MCP client (embedded in an AI application like Claude Desktop or an IDE) connects to one or more MCP servers. Each server exposes capabilities through three primitives:
- Tools — executable functions the AI can call (run a query, create a file, send a message)
- Resources — data the AI can read for context (files, database records, API responses)
- Prompts — reusable templates that guide AI behavior for specific tasks
All communication uses JSON-RPC messages. The client discovers what a server offers through capability negotiation at connection time — no hardcoded knowledge required.
Why it matters
Before MCP, every AI integration was bespoke. Connecting an AI to GitHub required GitHub-specific code. Connecting to Slack required different code. Connecting to a database required yet another integration. Each one was fragile and incompatible with other AI applications.
MCP solves this the way USB solved peripheral connectivity. One protocol, any AI, any tool. A server built once works with every MCP-compatible client. A client built once can use every MCP-compatible server. The ecosystem grows without coordination — server authors and client authors never need to talk to each other.
The protocol is maintained by Anthropic and published as an open specification.
See How MCP Works for the full architecture walkthrough.