MCP, the Model Context Protocol, is an open standard that connects AI assistants such as Claude and ChatGPT to outside tools and data. With MCP, an assistant can look up information and take actions in other apps, not only answer from what it already knows.
What MCP does
An AI assistant on its own knows only what it learned in training and what you type into the chat. MCP gives it a standard way to reach other systems, such as calendars, file stores, databases and workplace apps. The MCP website compares it to a USB-C port: one standard plug that works with many devices.
Because MCP is a shared standard, a tool's MCP server works with any AI assistant that can connect to MCP servers. The tool's maker builds one MCP server, and you choose which assistant to use it from.
How MCP works
An MCP connection has 2 sides. The AI assistant is the client. The outside system runs an MCP server. The client connects to the server, asks what the server offers, and uses those things when your request needs them.
An MCP server can offer 3 kinds of things:
- Tools: actions the assistant can take, such as booking a desk or searching a directory.
- Resources: data the assistant can read for context, such as file contents or database records.
- Prompts: reusable templates that help the assistant use the server well.
When you ask the assistant for something, it picks a tool, fills in the inputs, and sends the request to the server. The server does the work and sends back the result. The assistant then uses the result in its answer to you.
Local and remote MCP servers
A local MCP server runs on your own computer, next to the assistant. A remote MCP server runs on the internet, and the assistant reaches it at a web address. Robin uses a remote server.
A remote server can ask you to sign in before the assistant uses it. MCP uses OAuth for this, the same kind of sign-in you use for "Sign in with Google" or "Sign in with Microsoft". After you sign in, the assistant acts with your account and your permissions.
AI assistants that support MCP
Many AI assistants can connect to remote MCP servers, including Claude, ChatGPT, Gemini, Microsoft Copilot Studio, Grok and Perplexity. Each assistant uses its own name for an MCP connection, such as connectors, integrations, apps or MCP servers. Some assistants let only an admin or certain plans add one.
To learn more about the protocol, see the introduction to MCP and the MCP architecture overview on the MCP website.
How Robin uses MCP
Robin MCP is Robin's remote MCP server. It lets your AI assistant book desks and meeting rooms, look up people and buildings, and work with your floor plans for you. The Robin MCP site has the server address, setup steps for each assistant and the full list of actions. Its prompt library has example prompts to try.
Robin MCP offers tools, not resources or prompts. When you connect, you sign in with your Robin account, and each action runs as you. The assistant can see and do only what your Robin permissions allow. Robin receives the action and its inputs, not your conversation. Organizations hosted on Robin's EU infrastructure cannot connect AI assistants.
Connect your AI assistant to Robin
To start, see Using Robin in your AI assistant. It has the steps for each assistant. For Claude, see Connect Robin to Claude.
Admins control who can connect an AI assistant to Robin. To turn connections on or off, or to review the data and controls for a security review, see Security and governance for AI assistant connections.