Skip to main content
AI & Search

Model Context Protocol

MCP

Portrait of Lukas Horvath, co-founder of Roelu Studio
Lukas HorvathCo-founder

What is Model Context Protocol?

Model Context Protocol, or MCP, is an open standard introduced by Anthropic in late 2024 that defines how AI models connect to external tools, databases, and services. Instead of every AI app inventing its own integration pattern, MCP gives them one. A model that supports MCP can talk to any MCP-compliant server — your CRM, your file system, your analytics — without custom glue code on either side.

Why it matters

Before MCP, every AI integration was bespoke. Connecting Claude to your Notion was different from connecting ChatGPT to your Notion, and both broke when something changed. MCP is the standard that fixes that. By 2026, every major model supports it, and the ecosystem of MCP servers covers most of the tools your team already uses. For buyers evaluating AI products, MCP support is becoming the equivalent of REST API support a decade ago — table stakes. Vendors that do not support it are quietly being left out of agentic workflows.

How it works

MCP defines a small set of message types that pass between a client (the AI app) and a server (the tool or data source). The server exposes tools, resources, and prompts. The client calls them. Both sides speak JSON over standard transports like stdio or HTTP. A developer can spin up an MCP server in a few hundred lines of code, exposing anything from a database to a custom API. Once it is live, any MCP-compatible model can use it. That is the trick — the model does not need to be retrained for each tool, and the tool does not need to be rewritten for each model.

Related terms

  • AI Agent

    AI & Search

    An AI system that can take actions on its own — booking meetings, sending emails, querying databases, running code, updating records — instead of just…

  • Claude

    AI & Search

    Anthropic's AI assistant, known for careful reasoning, long context windows, and being a favorite among developers and writers who want a model that does not…

  • Large Language Model

    AI & Search

    A type of AI trained on huge volumes of text that can read, write, and answer questions in plain language — the engine behind ChatGPT, Claude, Gemini, and most…

  • Prompt Engineering

    AI & Search

    The practice of writing instructions to an AI model in a way that gets a reliable, useful result — part technical writing, part specification, part figuring…

  • Retrieval-Augmented Generation

    AI & Search

    A technique that lets an AI model look up fresh, specific information from a database or the web before answering, so it does not have to rely only on what it…

  • Composable Architecture

    Development

    An approach to building websites and digital products where each capability — CMS, search, payments, analytics — is a separate specialized service, connected…