Practical

Model Context Protocol (MCP)

Last reviewed: April 2026

A standardised open protocol that allows AI models to connect to external tools, data sources, and APIs through a universal interface.

The Model Context Protocol — MCP — is an open standard that defines how AI models communicate with external tools, databases, and services. Think of it as a universal adapter that lets any compatible AI assistant plug into any compatible data source without custom integration code.

The problem MCP solves

Before MCP, connecting an AI model to an external tool required bespoke integration work for every combination of model and tool. If you wanted Claude to query your CRM, you needed to build a specific connector. If you then wanted it to also check your project management tool, that was a second connector. Each integration was custom, fragile, and expensive to maintain.

MCP standardises these connections. A tool that implements the MCP server specification can be used by any AI model that supports the MCP client specification — no custom code required.

How MCP works

MCP uses a client-server architecture:

MCP server: Wraps an external tool or data source (a database, an API, a file system) and exposes its capabilities in a standardised format.
MCP client: Built into the AI application. It discovers available servers, understands what each can do, and calls them when the model determines it needs external information or wants to take an action.
Transport layer: Handles the actual communication between client and server, typically over standard protocols like HTTP or stdio.

Practical examples

An MCP server for your company's knowledge base lets any compatible AI assistant search and retrieve internal documents.
An MCP server for GitHub lets the AI read repositories, create issues, and open pull requests.
An MCP server for Slack lets the AI send messages and read channel history.

Why it matters for adoption

MCP dramatically lowers the cost and complexity of integrating AI into existing business systems. Instead of building and maintaining dozens of custom connectors, organisations can deploy MCP servers for their key systems once and make them available to any AI tool their teams use.

Current status

MCP was introduced by Anthropic in late 2024 and has seen rapid adoption. Major AI platforms and tool vendors are adding MCP support, and the ecosystem of pre-built MCP servers is growing quickly. It is becoming the de facto standard for AI-to-tool communication.

Want to go deeper?

This topic is covered in our Advanced level. Access all 100+ lessons free.

Why This Matters

MCP is transforming how organisations integrate AI into their existing technology stack. Understanding the protocol helps you evaluate which AI tools will connect most easily to your systems and plan an integration strategy that avoids vendor lock-in.

Related Terms

Tool Use (AI)

The ability of an AI model to call external tools, APIs, and services to perform actions or retrieve information beyond its built-in capabilities.

AI Agent

An AI system that can take actions autonomously — browsing the web, running code, calling APIs, and completing multi-step tasks with minimal human intervention.

API (Application Programming Interface)

A way for software to communicate with other software. APIs are how developers connect AI capabilities to websites, apps, and business tools.

Agentic Coding

Using AI agents to write, test, and debug code autonomously — planning multi-step development tasks and executing them with minimal human intervention.

Automation

Using technology to perform tasks without manual human effort. AI automation goes beyond traditional rule-based automation by handling unstructured tasks like writing, analysis, and decision-making.

Learn More

Continue learning in Advanced

This topic is covered in our lesson: Building Your Own AI Solutions

← Back to Glossary