Mailchimp Docs MCP Server

The Mailchimp Docs MCP Server connects an AI assistant to the developer documentation of the Mailchimp Marketing API. Anyone building an integration with Mailchimp, such as a newsletter sync, a sign-up flow, or a synchronization with a customer system, has to look things up constantly: which endpoint creates contacts, how do tags work, what do the error codes mean? With the server the assistant can answer these questions itself by searching the published documentation instead of guessing from memory. The server is operated by Mailchimp itself and described on the official developer site.

What is the Mailchimp Docs MCP Server?

MCP is an open standard through which an assistant calls the tools and knowledge sources of a service. Here the source is Mailchimp's own documentation site. According to the provider the server searches only the documentation and not your Mailchimp account. The endpoint is https://mailchimp.com/developer/_mcp/server, it is addressed as a hosted server over HTTP, and according to the page no Mailchimp sign-in is needed to search the documentation. So there is nothing to install and no key to enter.

What can it do, and what not?

This is the most important point before setup: the server reads, it does not act. It does not create campaigns, send emails, or change audiences. If you want an assistant to work inside your own account, the developer page points you to calling the API directly with an API key or OAuth 2. Separately from that, Mailchimp offers so-called Agent Skills, small guides as SKILL.md files, for example on correct consent handling, OAuth, webhooks, and bulk jobs. According to the page, documentation pages can also be fetched as Markdown by appending .md to the address, and there is an llms.txt as an index of all pages.

What is it particularly useful for?

The value lies in current and verifiable answers. An assistant without documentation access likes to invent parameters or use outdated endpoints. With the server it can read the matching page before writing code and rely on the current state of the documentation. Typical tasks are adding and updating contacts, working with segments and tags, setting up webhooks, or choosing between single and bulk endpoints. Because the server needs no credentials, it is also uncritical for teams that do not want to give their assistant access to customer data yet.

Setup

The page names four ways. In Claude Code a command of the form claude mcp add --transport http fern_mcp_mailchimp-docs https://mailchimp.com/developer/_mcp/server is enough. In Cursor you enter the address in the MCP configuration. In Claude Desktop you add an entry under mcpServers in the file claude_desktop_config.json, in VS Code an entry under servers in the file .vscode/mcp.json. Then restart the client; a connection indicator shows whether the server responds.

Security and limits

Since the server only delivers documentation and does not address an account, the risk is low. Still, note that search queries go to a hosted Mailchimp service; phrase them without customer data. Quality depends on the documentation: what is missing or outdated there, the assistant cannot output either. The server also does not replace a test against the real API. The page names no repository or license for the server, so we list none here.

Costs

According to the page, access to the documentation server is possible without a Mailchimp sign-in. Mailchimp's plans apply to using the Marketing API itself. According to the pricing page there is a free plan for small audiences with limited contacts and sending volumes, plus the paid Essentials, Standard, and Premium plans. Current limits are on Mailchimp's official pricing page.

Who is the Mailchimp Docs MCP Server for?

It is worthwhile for developers who integrate Mailchimp into a web application, a shop, or an internal system and work with an AI assistant while doing so. Anyone who only writes newsletters in the Mailchimp editor does not need it. Anyone who wants to automate tasks inside the account itself should know that this server is not meant for that and that the route leads through the API or the Agent Skills.

An MCP client that supports HTTP servers, such as Claude Code, Cursor, Claude Desktop, or VS Code. According to the provider, a Mailchimp account is not needed for searching the documentation.

The Mailchimp Docs MCP Server gives an AI assistant access to the developer documentation of the Marketing API. This guide shows setup in four common clients.

Step 1: Check the prerequisites

You need an MCP client that supports HTTP servers. According to the provider, neither a Mailchimp account nor an API key is needed for searching the documentation.

Step 2: Add it in Claude Code

Run the command claude mcp add --transport http fern_mcp_mailchimp-docs https://mailchimp.com/developer/_mcp/server in the terminal. Restart Claude Code afterwards.

claude mcp add --transport http fern_mcp_mailchimp-docs https://mailchimp.com/developer/_mcp/server

Enter it in Cursor, Claude Desktop, or VS Code

In Cursor add the address https://mailchimp.com/developer/_mcp/server in the MCP configuration. In Claude Desktop add an entry under mcpServers in the file claude_desktop_config.json, in VS Code an entry under servers in the file .vscode/mcp.json. Then restart the client.

Step 3: Check the connection

A connection indicator in the client shows whether the server responds. Then ask a concrete question, for example how to create a contact with consent, and make sure the assistant queries the documentation.

Know the limits

The server only searches the documentation and does not access your account. For tasks inside the account use the API with an API key or OAuth 2, or Mailchimp's separate Agent Skills. Always test generated code against the real API.

None. According to the provider, no Mailchimp sign-in is needed to search the documentation.

Read-only access to the public developer documentation; according to the provider the server does not access your Mailchimp account.

Search queries go to a hosted Mailchimp service. The server processes no account data. Documentation results enter the AI client context and therefore reach the model provider.

Low: search queries leave your machine, and outdated or missing documentation can lead to wrong answers from the assistant.