What is an MCP?
An open protocol that lets AI models connect to external tools, data sources, and services.
Stands for: Model Context Protocol
Pronunciation: /ˌem siː piː/
MCP explained in plain English
The Model Context Protocol (MCP) is a standardised way for AI applications to connect with external tools and data sources. Instead of building custom integrations for every tool, developers can use MCP to give AI models structured access to databases, APIs, file systems, and other services.\n\nMCP acts as a universal adapter between AI models and the outside world.
Analogy
Think of an AI model like a smartphone, and your data (like your calendar or files) like a pair of headphones. In the past, every AI needed a custom headphone jack to talk to a specific tool. MCP is like USB-C for AI — one universal standard port that lets any AI model instantly plug into any tool, calendar, or database without needing a custom-built connector every single time.
Example
Imagine a doctor's surgery website with an AI assistant chatbot. When a patient types: "Can I book an appointment with Dr. Smith this Thursday?" the AI uses an MCP connection to plug into the surgery's live calendar database. It checks the openings, finds a slot, and handles the booking automatically—all without a human receptionist needing to step in.
How is MCP used?
MCP is used to give AI chatbots and virtual assistants the "hands and feet" to interact with the real world. It allows an AI to securely read your local files, check your personal Google Calendar, look up data in a company database, or send a message on Slack, purely by translating human conversation into standard system actions.
Common misconceptions about MCP
MCP is relatively new and adoption is still growing. It is a protocol, not a product — the quality of integrations depends on individual MCP server implementations.
History
Related terms
People also read
- Agent
An AI system that can perceive its environment, make decisions, and take actions to achieve goals autonomously.
- act
A single action or decision made by an artificial intelligence system
- Backpropagation
The process that tells a neural network which internal settings caused an error and how to adjust them, working backwards through layers.
- feedback
Information provided about a process or system's performance
- GPT
A family of large language models developed by OpenAI that generate human-like text.
- LLM
A type of AI model trained on vast amounts of text to understand and generate human language.
- multi-agent collaboration
A process where multiple artificial intelligence agents work together to achieve a common goal
- RAG
A technique that combines AI language models with external knowledge retrieval for more accurate answers.
- reason
The ability of a system to draw inferences, make decisions, and solve problems
- Token
The basic unit of text that AI language models process, which may be a word, part of a word, or punctuation.