Model Context Protocol (MCP)¶
MCP is an open standard, introduced by anthropic in November 2024, for connecting AI applications to external systems — data sources, tools, and workflows. The canonical metaphor: "a USB-C port for AI applications" — one standardized way to plug any model into any data source or tool.
Why it matters¶
Before MCP, every AI-app ↔ tool integration was bespoke (N×M problem). MCP collapses that into a common protocol: build one MCP server for a data source, and any MCP-speaking client (claude-code, Claude, ChatGPT, IDEs…) can use it. This is foundational plumbing for the agent era and for owned infrastructure — you can run your own MCP servers exposing your own databases/tools to your own agents.
Key facts¶
- Architecture: MCP hosts (the AI app) run clients that connect to servers (which expose tools/data/prompts).
- Open-sourced; SDKs in TypeScript, Python, Java, Kotlin, C#, Go, PHP, Ruby, Rust, Swift.
- Broad adoption across major AI vendors and tools through 2025–2026.
- Transports:
stdio(local process),http,sse(remote). - Lets an agent access e.g. Google Calendar, Notion, a Postgres DB, or a Figma file through a uniform interface.
Related¶
- claude-code — an MCP client
- hermes — orchestrator that can consume MCP tools
- ai-agent-orchestration — MCP is the tool-access layer beneath orchestration