← All guides marginal.sh

An MCP Tool to Measure Email Engagement

Updated June 2026 · Hosted email MCP server

If you want an MCP tool to measure email engagement directly from an AI agent, marginal exposes managed sends plus open and click tracking so your assistant can read real results instead of guessing.

What 'measure email engagement' means here

Engagement is only useful if it's tied to a decision. marginal tracks opens and clicks per variant inside a single test, then reports lift between them so you can tell which subject line actually moved the needle.

Because everything runs through the email MCP server, the agent that wrote your subject lines is the same one that reads the numbers — no copy-pasting metrics out of a dashboard.

The tools that do the measuring

marginal ships four MCP tools. The measurement work mostly happens in get_results, but it only makes sense alongside the tools that create and ship the test.

Why agents are good at this

Measuring engagement is repetitive: send, wait, pull metrics, compare, decide. That loop maps cleanly onto AI agent email marketing workflows, where the agent can poll get_results and propose a follow-up without a human babysitting the funnel.

marginal is hosted, so there's no self-managed infrastructure. You point a supported client — Cursor, Claude Desktop, Claude Code, Windsurf, Cline, Continue, Codex, or Zed — at the endpoint and authenticate with a Bearer API key.

Getting started

Connect to https://marginal.sh/mcp, run your first subject line A/B test, and read engagement back through get_results. The free tier covers 100 experiments per month, which is plenty to validate the measurement loop before scaling up.

Full tool schemas and setup steps live in the docs at https://marginal.sh/docs/.

Get started with marginal

marginal is a hosted email marketing MCP server at marginal.sh. Sign up free, create an API key, and connect https://marginal.sh/mcp from Cursor, Codex, or Claude Desktop.

Documentation · Sign up · FAQ