Post-Purchase Subject Lines in Claude Desktop
Connect Claude Desktop to marginal's email MCP server and turn your post-purchase flow into a continuous subject line experiment — drafted, launched, and measured without leaving the chat.
Why post-purchase emails deserve their own subject line test
Post-purchase messages — order confirmations, shipping updates, review requests, and replenishment nudges — reach customers at a moment of high intent. A small open-rate gain on these sends compounds quickly because the audience already trusts you.
Running a subject line A/B test on this segment in Claude Desktop lets you experiment without spreadsheet juggling. You describe the email's purpose, marginal generates candidate subject lines, and you ship a measured test against your post-purchase list.
- Confirmation emails: test reassurance vs. excitement framing
- Review requests: test direct asks vs. incentive-led lines
- Replenishment reminders: test urgency vs. convenience angles
- Cross-sell follow-ups: test product-specific vs. curated phrasing
How marginal works with Claude Desktop
marginal is a hosted email MCP server, so Claude Desktop talks to it over the standard MCP protocol with a Bearer API key. There's nothing to self-host — you point Claude Desktop at the endpoint and the four tools become available in your conversation.
Each tool maps to a step in the experiment loop, so an AI agent email marketing workflow stays inside one chat thread.
- generate_variants — draft post-purchase subject line candidates
- launch_test — start the subject line A/B test and managed send
- get_results — pull open and click tracking with within-test lift
- recommend_next — get a data-backed suggestion for the next round
A sample post-purchase test in chat
Start by telling Claude Desktop the context: "Write five subject line variants for a post-purchase review request and run a subject line test on my post-purchase segment." Claude calls generate_variants, then launch_test, and reports back when results are in.
Once opens and clicks accumulate, ask for results. marginal returns the within-test lift between variants so you can see which line actually moved the needle, then recommend_next proposes how to iterate for the following send.
- Describe the email goal and audience in plain language
- Review and trim the generated variants before launching
- Check get_results once the send has had time to mature
- Use recommend_next to seed your next post-purchase test
Getting started and limits
The free tier includes 100 experiments per month, which is plenty for iterating on a handful of post-purchase touchpoints. Add the marginal endpoint at https://marginal.sh/mcp to Claude Desktop, authenticate with your marg_live_ API key, and you're ready to run your first subject line A/B test.
Full tool references and setup steps live in the docs at https://marginal.sh/docs/ if you want to script more advanced post-purchase flows.
- Endpoint: https://marginal.sh/mcp
- Auth: Bearer API key (marg_live_...)
- Free tier: 100 experiments/month
- Docs: 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.
- 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