Testing Post-Purchase Subject Lines in Cursor
Connect marginal to Cursor and let your AI agent draft, launch, and measure post-purchase subject lines without leaving the editor. This guide walks through wiring up the email MCP server and running your first subject line A/B test for order confirmations, shipping updates, and review requests.
Why post-purchase emails deserve a subject line test
Post-purchase is one of the highest-engagement moments in the customer lifecycle. Buyers already trust you, they're watching for their order, and open rates run well above promotional blasts. That makes the post-purchase sequence a low-risk, high-signal place to learn what subject line copy actually moves opens and clicks.
With marginal connected to Cursor, you can iterate on these subject lines as part of your normal dev workflow — ask the agent for variants, ship a subject line A/B test, and pull results back as structured data.
- Order confirmation: reassurance vs. upsell framing
- Shipping notification: tracking-first vs. excitement-first
- Delivery follow-up: review request vs. replenishment nudge
- Thank-you emails: personalization vs. brand voice
Connect marginal as an email MCP server in Cursor
marginal is a hosted MCP server, so there's nothing to self-host. Add the endpoint and your Bearer API key to Cursor's MCP configuration, and the four tools — generate_variants, launch_test, get_results, and recommend_next — become available to the agent.
Add the block below to your Cursor MCP settings and replace the placeholder with your key from https://marginal.sh/mcp.
- Endpoint: https://marginal.sh/mcp
- Auth: Bearer API key in the marg_live_... format
- Free tier covers 100 experiments per month
- Full reference at https://marginal.sh/docs/
Run your first post-purchase subject line test
Once connected, drive the whole flow with natural-language prompts in Cursor. Start by describing the post-purchase email and let the agent call generate_variants, then launch_test to send, and get_results to read open and click metrics with within-test lift.
A typical session looks like this:
- Prompt: "Generate 4 post-purchase subject lines for a shipping confirmation, balancing tracking info and brand tone."
- The agent calls generate_variants and proposes options to test
- Approve and ask it to launch_test against your post-purchase segment
- Later, ask for get_results to see opens, clicks, and lift per variant
- Use recommend_next so the agent suggests the follow-up subject line test
Make it part of AI agent email marketing
Because every step is exposed as an MCP tool, post-purchase subject line testing becomes a repeatable loop your agent runs for you. Cursor handles the prose, marginal handles the managed sends and tracking, and recommend_next keeps each new round informed by the last winner.
This is the practical shape of AI agent email marketing: you stay in the editor, describe the experiment, and the agent executes the subject line A/B test and reports lift — no context switching to a separate dashboard.
- Keep a running prompt template for each post-purchase touchpoint
- Promote winning variants and retest against new challengers
- Track lift over time to build a library of proven subject lines
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.
- 100 experiments/month on the free tier
- Four MCP tools: generate_variants, launch_test, get_results, recommend_next
- Listed in the MCP Registry as sh.marginal/mcp