WHITEBOXX / GLOSSARY
Model Context Protocol: one way
to connect AI to your systems.
The Model Context Protocol (MCP) is an open standard for connecting AI applications to outside tools and data, such as databases, file stores and business systems. A system that offers an MCP server can be used by any AI application that supports the protocol.
What is MCP?
Anthropic released MCP in November 2024. Before it, each AI application needed its own connector for each system. MCP defines one way for an application to find out what a system offers and to call it, much as any device that follows the USB standard works with any USB port.
In December 2025, Anthropic donated MCP to the Agentic AI Foundation, a fund under the Linux Foundation co-founded with Block and OpenAI. The protocol is now developed in the open.
HOW IT WORKS
The parts of MCP.
| Part | What it does |
|---|---|
| Host | The AI application the user works in, such as an assistant or an agent platform |
| Client | The connection the host opens to each server |
| Server | A program that offers a system’s capabilities to AI applications |
| Tools | Actions the model can call, such as running a query |
| Resources | Data the application can read, such as files or records |
| Prompts | Reusable instructions a server offers for common tasks |
What should you ask about MCP security?
An MCP server gives an AI application access to a system, so treat it like any other integration. Ask what it can read and change, whose permissions it uses, how it is authenticated and whether its actions are logged.
Text returned by a server can contain instructions aimed at the model, which is a form of prompt injection. Connect only servers you trust, and limit each one to the tools a task needs.
How Evidence OS connects to your systems.
Our engineers set up the connections to your business systems for each project, with access agreed with your team. Each Evidence OS run records the data it used.
See platform integrations