Model Context Protocol (MCP)
An open standard that lets AI models connect to external tools, apps, and data sources in a consistent way.
What is Model Context Protocol (MCP)?
MCP is an open standard that lets an AI assistant connect to outside tools, files, and apps in one consistent way, instead of needing separate custom-built code for every single connection. Think of it like a USB-C port for AI: any tool that speaks MCP can plug into any AI application that also speaks MCP, without a bespoke integration for each pairing. For example, a company can build one MCP server for its internal database, and any MCP-compatible AI assistant, from any vendor, can then query that database the same way.
Why it matters
Anthropic open-sourced MCP on November 25, 2024, defining a client-server architecture in which "MCP servers" expose resources, tools, and prompts over JSON-RPC 2.0, and "MCP hosts" (AI applications) connect to them through clients. Adoption accelerated through 2025: OpenAI, Google DeepMind, and Microsoft all added MCP support to their products, and Microsoft and GitHub joined an MCP steering committee that May. On December 9, 2025, Anthropic transferred stewardship of MCP to the newly formed Agentic AI Foundation (AAIF), a directed fund under the Linux Foundation co-founded by Anthropic, Block, and OpenAI, with backing from Google, Microsoft, AWS, Cloudflare, and Bloomberg, making MCP a vendor-neutral standard governed outside any single company. By the time of the donation, more than 10,000 active public MCP servers existed, spanning developer tools to enterprise deployments.