Verified

Mailgun MCP Server

Mailgun email operations and deliverability analytics managed via AI agents.

Data last scanned today · Reviewed today

4515 tools★ 60TypeScriptApache-2.0

Overview

This MCP server provides a functional interface for managing Mailgun account workflows directly through AI assistants. It enables capabilities including sending emails, managing domain configurations, and overseeing inbound email routes and mailing lists. The server operates locally over stdio and utilizes an OpenAPI-driven architecture to register Mailgun endpoints as executable tools. Beyond basic message handling, it supports advanced deliverability tasks such as validating email addresses, inspecting email previews, and retrieving inbox placement test results. Developers can narrow the toolset using tag filtering at startup to limit the exposure to specific products like validation, optimization, or inspection tools.

Our verdict

With a security score of 40/100, this server is classified as high-risk. This rating reflects its community-maintained status rather than being an official vendor release, and the fact that its authentication methods and read-only mode claims have not undergone independent editorial review. Because this tool handles sensitive email account operations and infrastructure, users should be cautious and audit the implementation before deploying it in environments requiring strict security guarantees.

Uses Zod for robust input schema generation
Limits blast radius by excluding delete operations
Maintains a small footprint with only three dependencies
Includes tag filtering to limit accessible toolsets

Not an official release from Mailgun
Authentication methods remain unreviewed
Read-only mode guarantees are unverified

Setup

Claude Desktop

{
  "mcpServers": {
    "mailgun-mcp-server": {
      "args": [
        "-y",
        "@mailgun/mcp-server"
      ],
      "command": "npx"
    }
  }
}

Paste into ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows), then restart Claude Desktop. Some servers require API keys or tokens as environment variables — check the project's README and add an "env" object if needed.

No read-only mode — running this grants full read/write access.
No built-in authentication — anyone with access to this config can invoke every tool it exposes.

Tools

Tool	Description	Risk
send_email	Send emails, retrieve stored messages, resend messages.	medium
manage_domains	View domain details, verify DNS configuration, manage tracking settings.	medium
manage_webhooks	List, create, and update event webhooks.	medium
manage_routes	View and update inbound email routing rules.	medium
manage_mailing_lists	Create, view, and update mailing lists and their members.	medium
manage_templates	Create, view, and update email templates with versioning.	medium
get_analytics	Query sending metrics, usage metrics, and logs.	low
get_stats	View aggregate statistics by domain, tag, provider, device, and country.	low
get_suppressions	View bounces, unsubscribes, complaints, and allowlist entries.	low
get_ips_and_pools	View IP assignments and dedicated IP pool configuration.	low

Compatibility

Client	Local	Docker	Remote	Read-only
Claude Desktop
Cursor
VS Code
Windsurf
ChatGPT

Frequently asked questions

›What are the system requirements for running this server?

You need Node.js version 20.12 or higher and a valid Mailgun account with an API key.

›Does this server support Mailgun accounts hosted in the EU region?

Yes, you can configure the server for the EU region by setting the MAILGUN_API_REGION environment variable to "eu". It defaults to "us" if not specified.

›Can this MCP server delete data from my Mailgun account?

No, the tools are limited to read and update operations only. No delete operations are exposed to ensure the blast radius of unintended actions remains small.

›Is there a hosted version of this server I can connect to?

No, Mailgun does not currently offer a hosted version. This MCP server must be run locally on your machine where it communicates over stdio.

›How can I limit the tools exposed to my AI agent?

You can use tag filtering at startup to scope which tools are registered. Supported tags include send, validate, optimize, and inspect.