Aurora SendCloud MCP Server

Overview

A Model Context Protocol (MCP) server for Aurora SendCloud, enabling MCP-compatible AI clients to send emails and query deliverability analytics.

Note: This MCP server runs locally on your machine. Aurora SendCloud does not currently provide a hosted MCP server.

What can it do?

This MCP server keeps the tool surface small (send_email, get_email_status, get_send_statistics, get_account_quota) but each tool supports multiple operation modes and richer analysis output.

Capabilities

• Messaging - Send transactional emails
• Message Tracking - Query per-message delivery status
• Statistics - Query daily/hourly/invalid classification metrics by period or date range
• Account Quota - Query account daily quota and usage

Prerequisites

• Node.js (v18.17 or higher)
• Aurora SendCloud account (API_USER + API_KEY)

Quick Start

Configuration

Add the following to your MCP client configuration:

{
  "mcpServers": {
    "aurorasendcloud": {
      "command": "npx",
      "args": ["-y", "@aurorasendcloud/mcp-server"],
      "env": {
        "SENDCLOUD_API_USER": "YOUR_API_USER",
        "SENDCLOUD_API_KEY": "YOUR_API_KEY",
        "SENDCLOUD_REGION": "sg"
      }
    }
  }
}

Environment Variables

| Variable | Required | Default | Description | |---|---|---|---| | SENDCLOUD_API_USER | Yes | N/A | Aurora SendCloud API_USER | | SENDCLOUD_API_KEY | Yes | N/A | Aurora SendCloud API_KEY | | SENDCLOUD_REGION | No | sg | Region: sg / us / hk | | SENDCLOUD_BASE_URL | No | N/A | Custom API base URL (highest priority) |

Region Mapping

• sg => https://api.aurorasendcloud.com
• us => https://api-us.aurorasendcloud.com
• hk => https://api-hk.aurorasendcloud.com

Client-Specific Config Paths

• Claude Desktop (macOS): ~/Library/Application Support/Claude/claude_desktop_config.json
• Claude Desktop (Windows): %APPDATA%/Claude/claude_desktop_config.json
• Claude Code: Run claude mcp add or edit ~/.claude.json

Tools

• send_email
• get_email_status
• get_send_statistics
• get_account_quota

What does it look like in action?

Typical workflow in an MCP client:

1. Send a campaign or template email with send_email.
2. Track delivery/opens/clicks with get_email_status (include_tracking=true).
3. Pull daily or hourly trends with get_send_statistics.
4. Check quota risk with get_account_quota (include_health=true).

Sample Prompts

Send an Email

Send an email from [email protected] to [email protected] with subject "Welcome" and a short HTML body.

Fetch Statistics

Get send statistics for the last 7 days for domain example.com and summarize delivery, open, and bounce trends.

Compare Domains

Get last month statistics grouped by domain and compare deliveredNum and bounceNum.

Check Quota

Check my account quota and tell me if I'm at risk of hitting limits today.

Development

To run from source:

git clone <your-repository-url>
cd aurorasendcloud-mcp
npm install
npm run build
npm run start

In MCP client config, use:

{
  "mcpServers": {
    "aurorasendcloud": {
      "command": "node",
      "args": ["/path/to/aurorasendcloud-mcp/dist/index.js"],
      "env": {
        "SENDCLOUD_API_USER": "YOUR_API_USER",
        "SENDCLOUD_API_KEY": "YOUR_API_KEY"
      }
    }
  }
}

Security Considerations

API credential isolation

Credentials are provided through environment variables or per-tool inputs and only used by the local MCP server process for Aurora API requests.

Local execution

The server runs locally and communicates with Aurora APIs over HTTPS.

Least-privilege credentials

Use dedicated API credentials with minimal required permissions.

Prompt injection

As with any MCP server, untrusted prompts could induce unintended tool calls. Review tool call confirmations in your MCP client.

Debugging

The MCP server communicates over stdio. Refer to the MCP Debugging Guide.

License

ISC