MCP server

The Model Context Protocol (MCP) lets an AI coding agent such as Claude Code, Cursor or Copilot call tools while it works. The api.email MCP server gives your agent the same capabilities as the API, so it can send a test email, check a domain or read an inbound message while it writes your integration.

What the agent can do

The server exposes a small set of tools that map one to one to API endpoints:

• send_email sends an email with the same fields as POST /v1/emails.
• get_email returns the status and events of an email by id.
• list_domains shows your domains and the DNS records each one still needs.
• list_inbound returns recent email.received events for your routes.

Because every tool is a thin wrapper over the public API, anything the agent does is visible in your dashboard and logs, exactly like requests from your own code.

Connect an agent

Add the server to your agent’s MCP configuration with the server URL https://api.email/mcp and an API key in the Authorization header. Use a test key while the agent is exploring: tool calls behave normally, but no email reaches real recipients. Switch to a live key only for tasks you have reviewed.

Docs for agents

Agents work best with plain text. The full documentation is available as llms.txt at the site root, and curl api.email prints a short usage guide right in the terminal. Point your agent at either when you ask it to add sending or receiving to your app.

Safety notes

Treat an agent like any other client: give it its own key, scope it to the environment it works in, and revoke the key when the task is done. Review the code it writes before deploying it, especially anything that sends email to real users.