MCP (Model Context Protocol)
A standard protocol for connecting AI assistants to external tools and data sources.
Simple Definition
MCP (Model Context Protocol) is a standardized way to connect AI assistants to external tools, databases, and services. Think of it like USB for AI—a universal connector that lets different AI tools talk to different data sources.
Technical Definition
MCP is an open protocol (created by Anthropic, late 2024) that defines:
- Resources: Data that AI can read (files, databases, APIs)
- Tools: Actions AI can take (run commands, query databases)
- Prompts: Reusable prompt templates
MCP servers expose capabilities; MCP clients (AI assistants) consume them.
[AI Assistant] ←MCP→ [MCP Server] ←→ [Database/API/Files]
Why It Matters
Before MCP: Every AI tool built custom integrations. If you wanted Cursor to access your database, you built a Cursor-specific integration. Want Claude Desktop to access the same database? Build another integration.
With MCP: Build one MCP server for your database. Any MCP-compatible client (Cursor, Claude Desktop, Cline, etc.) can use it.
Adoption (2025): MCP is supported by Cursor, Claude Desktop, Cline, and others. Growing ecosystem of community-built MCP servers.
Key Takeaways
- MCP is a universal protocol for AI-to-tool connections
- Created by Anthropic in late 2024
- Enables one integration to work with multiple AI clients
- Supported by Cursor, Claude Desktop, Cline, and others