What is Model Context Protocol (MCP)?
Model Context Protocol: An open standard for connecting AI apps and agents to tools and data through MCP servers, so one integration works across many AI clients.
The Model Context Protocol is an open protocol, introduced by Anthropic in late 2024, that standardizes how AI applications connect to external systems. An MCP server wraps a system (a database, a SaaS API, a file store) and exposes it as tools the model can call, resources it can read, and prompt templates. An MCP client, such as Claude, ChatGPT, Cursor, or VS Code, discovers those capabilities and lets the model use them. In December 2025 Anthropic donated MCP to the Agentic AI Foundation under the Linux Foundation, and it is now supported by most major AI clients.
MCP replaces one-off integrations with a shared interface: build a server once and any compatible client can use it. Servers can run locally or remotely over HTTP, with OAuth for authorization. The specification keeps evolving, with 2026 work focused on stateless operation for easier scaling, interactive UI extensions, and enterprise needs like audit trails and gateways.
Why it matters when you build or buy AI
If you want your own systems available to AI assistants your team already uses, an MCP server is now the standard way to do it. It is also a security surface. Each server's tools act with whatever credentials you give them, and malicious or careless tool descriptions can steer a model ("tool poisoning"). Only install servers you trust, scope permissions tightly, and log every tool call.
Example
A company builds an MCP server for its internal order system, so staff can ask Claude "which wholesale orders shipped late this week" and get answers from live data.
Related terms
More in Agents & Automation
Book 20 minutes, or send a brief
A former CTO replies. You leave with a written plan and a fixed estimate.
Prefer LinkedIn? Connect with Harish.