← All guides marginal.sh

Testing Re-engagement Subject Lines in Cursor

Updated June 2026 · Hosted email MCP server

Win back dormant subscribers without leaving your editor. With marginal's email MCP server connected to Cursor, you can generate re-engagement subject lines, launch an A/B test, and read open and click lift directly from your AI agent chat.

Why run re-engagement tests from Cursor

Re-engagement campaigns live or die by the subject line — it's the only thing a lapsed subscriber sees before deciding to open or ignore. Cursor's AI agent gives you a fast loop for drafting, testing, and iterating on those lines while you're already in a coding context, so you don't have to context-switch into a separate marketing dashboard.

Because marginal is a hosted email MCP server, Cursor talks to it over a single endpoint with a Bearer API key. There's no self-hosted infrastructure to maintain and no separate ESP keys to wire up before you can launch your first subject line test.

The re-engagement subject line workflow

A re-engagement subject line A/B test in Cursor follows a simple chain of marginal's MCP tools. You ask the agent for variants, pick a few, launch a managed send to your dormant list, then come back for results once opens accumulate.

Writing re-engagement variants that earn opens

When you prompt generate_variants in Cursor, give it the context that matters for a re-engagement audience: how long they've been inactive, what they originally signed up for, and any incentive you're offering. The more specific the brief, the sharper the subject line test.

Try contrasting distinct strategies in a single test rather than minor word tweaks. Curiosity, a direct "still want these emails?" prompt, and an incentive-led line will tell you far more about your lapsed segment than three near-identical phrasings.

Reading lift and deciding the winner

After launch_test runs, call get_results in Cursor to see open and click tracking per variant alongside within-test lift metrics. Lift tells you how each subject line performed relative to the others in the same send, which is what you want when judging re-engagement performance.

Use recommend_next to have the AI agent interpret the numbers and propose your next move — promoting the winner to your broader inactive list or queuing a follow-up test on a new angle. On the free tier you get 100 experiments per month, which is plenty of room to iterate on win-back copy. See the docs at https://marginal.sh/docs/ or the endpoint at https://marginal.sh/mcp to connect Cursor.

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