Definition
An open standard, introduced by Anthropic in November 2024, that lets AI assistants such as Claude connect to outside tools and data through servers. One MCP server makes an app’s actions available to any compatible assistant.
A common plug for AI assistants
Claude can write a good follow-up, but without access to your tools it can’t tell which leads need one. The Model Context Protocol (MCP) gives AI assistants a common plug to reach outside tools and data. Anthropic introduced it as an open standard in November 2024. In December 2025, it was handed to the Agentic AI Foundation, under the Linux Foundation, and many AI assistants and developer tools now support it.
How it works
- An MCP server is published by an app. It describes the actions the app offers, called tools, and the data it can share.
- An MCP client lives in the assistant, such as Claude. It connects to the server and shows the model what is available.
- When you ask for something, the model picks the right tools, the client calls them with your permission, and the result comes back in the conversation.
Compared with a classic integration, nobody writes code for each assistant: one server works with every compatible client. It often sits on top of the app’s existing API.
What it changes for a sales team
Questions that used to need an export and a spreadsheet become a sentence: “Which leads qualified this week but have no appointment?”, “Draft a follow-up for the ones who asked about financing.” The assistant reads live data through the server and acts within the permissions you gave it. The usual care applies: connect only servers you trust, give them the access they need and no more, and review any action that sends messages or changes records.
WhatSetter’s MCP server
WhatSetter has an MCP server: you can drive WhatSetter from Claude, on your own account. It sits next to the REST API, which covers leads, conversations, lists, sends, prompts, appointments, groups and knowledge bases, and next to the signed webhooks that tell your systems when something happens. Together, they let a developer, or an assistant, read and act on your WhatSetter account from outside the dashboard. See integrations & API.