Mxmcp Py

🚀 MCP Server for mxHERO Multi-Account Email Search

This MCP (model context protocol) server, a Python project, provides access to mxHERO's Mail2Cloud Advanced multi-account email search service. The Model Context Protocol (MCP) standardizes how models interact with diverse data sources and services. In this project, MCP enables seamless integration with mxHERO Mail2Cloud Advanced, a high - performance data service for company email data. It connects to company email services and optimizes content for fast, scalable, and secure access by AI solutions.

🚀 Quick Start

To quickly start exploring this MCP server, follow the installation and usage steps below.

✨ Features

Architecture

Mail2Cloud selectively captures emails from one or more accounts. Powerful filters can precisely control email selection by examining any aspect of messages and their attachments. Captured emails are optimized and stored in an isolated tenant in a vector database for email - related searches. This MCP accesses the stored emails in the tenant via authenticated access credentials.

Advantages

Solutions built with Mail2Cloud Advanced MCP outperform other AI solutions in email data search and knowledge recovery (study).

• It provides secure links to original emails, protecting them from accidental user deletion.
• It allows LLMs to search large - scale email repositories, beyond their context window limitations.

Demo Accounts

mxHERO offers demo accounts pre - loaded with thousands of emails to help you explore this MCP. More details about the demo emails can be found here.

See 'Access Tokens' below to get a token.

For more information, visit mxHERO Mail2Cloud Advanced multi - email account service, including architecture and optimizations.

📦 Installation

1. Clone the repository

git clone https://github.com/mxaiorg/mxmcp-py

2. Install uv if not installed yet:
   • MacOS+Linux

   curl -LsSf https://astral.sh/uv/install.sh | sh
   

   • Windows

   powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"
   

3. Create and set up our project:
   • MacOS+Linux

   # Create virtual environment and activate it
   uv venv
   source .venv/bin/activate
   
   # Lock the dependencies
   uv lock
   

   • Windows

   # Create virtual environment and activate it
   uv venv
   .venv\Scripts\activate
   
   # Lock the dependencies
   uv lock
   

💻 Usage Examples

Running the MCP Server

To start the MCP server, run the following command:

uv --directory /fullpath/PycharmProjects/mxmcp-py run src/mxmcp.py --token "my_token"

Normally, your client will start the server, but this command helps you ensure everything works.

Using Claude as the client

1. Edit your claude_desktop_config.json:

code ~/Library/Application\ Support/Claude/claude_desktop_config.json

You can also use your preferred JSON editor. You may need to create the file if it doesn't exist. 2. Add the configuration: Replace the --directory argument value ('/Users/...') with the absolute path to the python script.

{
    "mcpServers": {
        "mxhero-mcp-server": {
            "command": "uv",
            "args": [
                "--directory",
                "/Users/your_user/Desktop/mxmcp-py",
                "run",
                "src/mxmcp.py",
                "--token",
                "<copy_your_token_here>"
            ]
        }
    }
}

📚 Documentation

Alternate versions

A Go version (plus prebuilt binaries) can be found here.

Streamable HTTP

This MCP repo is the 'stdio' variant. HTTP options are available at the following addresses:

• https://lab4-api.mxhero.com/mcp/connect (streamable HTTP)
• https://lab4-api.mxhero.com/mcp/sse (Legacy SSE)

Tools implemented

email_search

Search stored emails Parameters

• query (str): Email search query Returns JSON of search results

Requirements

• Python 3.13 or higher
• mxHERO Vector Search token

Access Tokens

• A demo token can be obtained at https://lab4-api.mxhero.com/demo_signup
• For production tokens, uncheck "Demo" and Request an account (or contact mxHERO at contact@mxhero.com).

MCP library

https://gofastmcp.com/servers/tools