Skip to main content

AI glossary Products & Platforms

What is the Model Context Protocol (MCP)?

MCP (Model Context Protocol)

← All glossary terms
In short

An open standard that defines how AI applications connect to tools, data and enterprise systems through a common client-server protocol.

Explained

The Model Context Protocol (MCP) is an open standard, introduced by Anthropic, that defines how an AI application discovers and uses external tools, data sources and prompts. Instead of writing a bespoke integration for every model and every system, a team wraps a system once as an MCP server, and any MCP-compatible application can use it. The protocol follows a client-server model: the host application opens one client per server, and each server exposes what it offers in a structured, self-describing way. This reduces integration work and makes agent architectures more composable, but it does not remove the need for permissions, authentication and governance.

How MCP works

MCP has three roles. The host is the AI application the user works with, such as a chat assistant, an IDE or an agent runtime. Inside the host, a client keeps a connection to one MCP server. The server wraps a system (a database, a document store, a SaaS tool, an internal API) and describes what it offers.

Servers expose three kinds of primitive:

  • Tools: actions the model can ask to run, each with a name, a description and an input schema.
  • Resources: data the application can read and place in the model's context, such as files or records.
  • Prompts: reusable templates that the server offers to the user or the application.

Local servers usually talk to the host over standard input and output; remote servers use HTTP. The model never calls the system directly: it proposes a tool call, the host executes it through the client and returns the result.

MCP vs a regular API integration

An API is the interface of one system. An MCP server is an adapter that presents that interface in a form any AI application can discover and use without custom code. The API still exists underneath; MCP standardises the layer between the model and the API.

The practical difference is reuse. Without MCP, connecting five systems to three AI applications means fifteen integrations. With MCP, it means five servers that every compatible application can use. The trade-off is one more component to run, secure and version.

When it makes sense, and what to watch

  • Several AI applications or agents need the same internal systems.
  • You want to swap models or AI vendors without rewriting integrations.
  • Tools change often and you want them described in one place.

Watch the security side: an MCP server acts with whatever credentials it is given, so scope them tightly, check who can install servers, and treat tool descriptions and returned content from third-party servers as untrusted input, because they can carry prompt injection.

Frequently asked questions

What does MCP stand for in AI?

MCP stands for Model Context Protocol. It is an open protocol that standardises how AI applications connect to external tools, data and prompts, so that one integration can serve many applications.

Who created the Model Context Protocol?

Anthropic introduced MCP as an open standard and published its specification and SDKs as open source. It is now supported by many AI applications and developer tools beyond Anthropic's own products.

Is MCP the same as an API?

No. An API is the interface of a specific system; MCP is a protocol for exposing such interfaces to AI applications in a common, self-describing way. An MCP server usually calls an existing API underneath.

What is the difference between MCP and tool calling?

Tool calling is the model's ability to emit a structured request to run a function. MCP is the standard way to make those functions available to an application: servers describe their tools, and the host passes them to the model and executes the calls.

Is MCP secure?

MCP itself is a protocol; security depends on how servers are built and deployed. Give each server the minimum permissions it needs, authenticate remote servers, log tool calls and review third-party servers before installing them.

Decision guides that use this term

Where it shows up at Thinkia

Products

Further reading

Sources

Updated: 2026-10-21

Want to see how this applies to your organisation?

Talk to an AI Expert