Test Seasonal Promotion Subject Lines in Zed
Connect Zed to marginal's email MCP server and let your AI agent draft, test, and measure seasonal promotion subject lines without leaving your editor.
Why run seasonal subject line tests from Zed?
Seasonal promotions live and die on timing. A holiday sale, an end-of-summer clearance, or a back-to-school push only has a short window to land in the inbox and earn the open. Testing subject lines manually across that window is slow, and by the time you've picked a winner the season has moved on.
Zed's agentic workflow lets you keep the campaign work next to the rest of your code and copy. With marginal wired in as an email MCP server, your AI agent can generate seasonal promotion subject lines, launch a subject line A/B test, and read back results — all from the same chat panel you already use to ship features.
- Iterate on seasonal copy fast, while the promotion is still live
- Keep campaign logic versioned alongside your project context
- Hand the agent a brief and get measurable lift, not guesses
Generating seasonal promotion variants
Start by describing the promotion to your Zed agent: the offer, the season, the audience, and the tone you want. The agent calls marginal's generate_variants tool to produce a spread of subject lines — some urgency-driven, some curiosity-driven, some leaning on the seasonal hook itself.
Good seasonal subject lines tend to mix the timely angle with a concrete benefit. Ask the agent to vary that balance so your subject line test covers a meaningful range instead of near-duplicates.
- Lead with the seasonal moment: "Last day of our Spring Sale"
- Pair the season with the discount: "30% off, just in time for winter"
- Test emoji vs. plain text for festive promotions
- Try a curiosity angle that hints at the offer without stating it
Launching the A/B test and reading lift
Once you've picked the variants worth testing, the agent uses launch_test to send the seasonal promotion as a managed send through marginal. Opens and clicks are tracked automatically, and get_results returns within-test lift so you can see which subject line is pulling ahead.
Because seasonal campaigns are time-sensitive, lean on recommend_next: it tells the agent whether you have a clear winner yet or whether the difference is still inside the noise. That keeps you from declaring a winner too early on a short-lived promotion.
- launch_test — send the seasonal promotion and split traffic across variants
- get_results — track opens, clicks, and within-test lift
- recommend_next — decide when to roll the winner to the full list
Connecting marginal in Zed
marginal is a hosted MCP server, so there's nothing to deploy. Point Zed at the endpoint https://marginal.sh/mcp and authenticate with your Bearer API key (the marg_live_... key from your account). The free tier covers 100 experiments per month — plenty for testing seasonal promotion subject lines across a few campaigns.
Full setup and tool details are in the docs at https://marginal.sh/docs/. Once connected, your AI agent email marketing workflow runs end to end inside Zed.
- Endpoint: https://marginal.sh/mcp
- Auth: Bearer API key (marg_live_...)
- Registry: sh.marginal/mcp
- 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