MCP

Open protocol that standardizes how AI applications connect to external systems. Enables LLM-powered tools to reach data sources such as local files and databases, tools such as search and calculators, and templated workflows. Think of it as a USB-C port for AI: a single standard for plugging applications into context and capabilities.

Adopted by AI assistants such as Claude and ChatGPT, development tools such as VS Code and Cursor, and many other clients. Inspired by the Language Server Protocol, it brings the same ecosystem approach to AI integrations. Servers expose resources, prompts, and tools over JSON-RPC; clients can offer sampling, roots, and elicitation. Build once, integrate across many AI applications.

Key capabilities:

• JSON-RPC 2.0 message format with stateful connections and capability negotiation
• Servers expose resources (context and data), prompts (templated workflows), and tools (functions for models to execute)
• Clients can offer sampling, roots, and elicitation for server-initiated behaviors
• Standardized protocol reduces development time when building or integrating with AI applications
• Broad ecosystem support across AI assistants and development tools

Developers use MCP to expose proprietary data or internal tools to AI assistants without custom integrations per client. IDE extensions give coding assistants access to project files, databases, or API documentation through a single protocol. Enterprise teams connect chatbots to multiple data sources so users can analyze data via chat. The protocol enables agents to act on calendars, design tools, and workflows from a unified interface.