AI Email Marketing Copy Generator
The highest-ROI email copy isn't clever, it's specific. A good abandoned-cart email names the actual product the shopper left. A good welcome sequence knows whether the signup was an email-gate for a coupon or a full-funnel opt-in. scrb generates the common e-commerce email flows - abandoned cart, welcome series, post-purchase, broadcast, win-back - as structured copy (subject + preview text + body) that you paste into Klaviyo, Mailchimp, Customer.io, ConvertKit, or whatever your ESP is. Spam-trigger words filtered out by default; 25+ languages; 5-700-word range per email.
What scrb knows about email marketing that a chat model doesn't
Chat models treat email as "write me some text about X". Scrb knows the specific archetypes and their structural rules.
- Five archetypes: abandoned cart, welcome series (3-5 emails), post-purchase, broadcast (promo / product launch), win-back. Each has a different JSON shape and different subject-line discipline.
- Subject line + preview text: subject capped at 40-50 chars (mobile inbox visibility), preview text 85-100 chars (fills the subject-line second row). Both generated per-email and scored for spam-trigger words before return.
- Spam-word filtering: Gmail's promo-tab triggers - "FREE!!!", "act now", "limited time", "100% guaranteed", excessive punctuation, all-caps subject - are banned from output. Output still converts, just without tripping Gmail's classifier.
- Welcome series as a sequence, not single emails: request a "welcome_series" and scrb returns 3-5 emails with deliberate spacing (Day 0 / Day 1 / Day 3 / Day 7 / Day 14), each building on the last rather than re-introducing the brand every time.
- Abandoned cart knows the product: supply the product name + price + cart total and the email body references them specifically instead of a generic "you left something in your cart" line.
- Plain text + HTML-ready: scrb returns body as plain Markdown-like text that renders cleanly in any ESP's rich-text editor. No injected HTML classes that your ESP will strip on save.
- Tone calibrated per archetype: win-back is apologetic and brief; post-purchase is warm and future-looking; broadcast is concrete (price + deadline + CTA); welcome series is progressive (Day 0 = thanks, Day 3 = feature deep-dive, Day 7 = social proof).
- 25+ languages native: generate the same sequence directly in French, German, Japanese, Portuguese. Idiomatic, not translated.
Before / after: real examples
Generic AI output:
Subject: Don't forget! Your cart is waiting 🛒✨scrb output:
Body: Hey! We noticed you left some amazing items in your cart. Don't miss out on your exclusive products! Act now before they sell out! Complete your purchase today to unlock your potential...
Subject (41 chars): Did the charcoal beanie not quite fit?Why this wins: the subject is a question, not a shouty exclamation. It references the specific product. It respects the reader ("we'd actually want to know" invites a reply, which signals to Gmail the brand is a genuine sender). Spam triggers ("act now", "don't miss out", "exclusive products", "unlock") are gone.
Preview (89 chars): It's still in your cart. Free shipping through Friday if you want to give it another look.
Body:
Hey,
You were looking at the Merino Wool Beanie in charcoal a couple of days ago. It's still sitting in your cart - we didn't clear it.
Not sure if you got distracted, or if you saw something that made you pause. If it's the former, here's a direct link back to your cart:
[Return to cart → $64]
If it's the latter - size, colour, or just changed your mind - we'd actually want to know. Hit reply and tell us. We're a small team; the email lands in an actual human inbox.
Free shipping through Friday if the beanie is still tempting you.
- The team at [brand]
P.S. If you want to clear the cart and start fresh, unsubscribe from cart reminders.
scrb output (subjects + Day):
Day 0 (41 chars): Your 10% code: SKIN10. And a question.Why this wins: subjects are short enough for mobile Gmail, question-framed rather than exclamatory, and the series progresses from "here's your code" through teaching and social proof to a clear deadline - not four variations of "don't forget about us".
Day 1 (38 chars): The one ingredient we won't use.
Day 3 (44 chars): What our most returning customers buy first
Day 7 (33 chars): Your code expires Friday.
Progression logic:
• Day 0: deliver the code + ask one question ("What's the skincare problem you hope we solve?")
• Day 1: teach something concrete, build trust (one ingredient the brand won't use, with a specific reason why)
• Day 3: social proof via a specific pattern from data (not "loved by thousands"), framed as a purchase recommendation
• Day 7: last-call reminder with the expiry, single CTA, no padding
FAQ
Which email archetypes does scrb generate?
Five: abandoned_cart (single email), welcome_series (3-5 emails), post_purchase (single), broadcast (promo / product launch, single), win_back (single or 2-email sequence). Each has a specific JSON shape and tone calibration. Request one at a time via the API, or select from the dropdown in the web app.
Does scrb integrate with my ESP?
Not as a native plugin yet - the output is plain text + structured JSON that you paste into your ESP's editor. Klaviyo, Mailchimp, Customer.io, ConvertKit, and Flodesk all accept the output without modification. Native plugin candidates are on the roadmap based on customer demand; the Klaviyo plugin is the likeliest first.
Are the subject lines tested against spam filters?
scrb runs the output through a spam-trigger heuristic (banned substrings + punctuation + all-caps check) before returning. It's not a full spam-score like Mail-Tester, but it catches the top ~80% of triggers before they hit your ESP. For critical broadcasts to large lists (>10k), still run through Mail-Tester before sending.
How does billing work for a welcome series?
A 3-email welcome series = 3 generations debited. A 5-email series = 5. Quota is checked upfront so you can't accidentally get a partial series generated.
Can I localize a series into multiple languages?
Yes - request the same series in each target language. Each language call debits its own quota. For a DTC brand running EN + FR + DE, a 4-email welcome series × 3 languages = 12 generations. Each language version is written idiomatically, not translated.
Does scrb personalize with merge tags?
Yes - you can set merge_tags like {first_name} or {cart_total} in the input, and scrb leaves them as placeholders in the output so your ESP can fill them. Standard Liquid / Handlebars syntax works.
How to plug scrb into your workflow
- Web app: pick an email archetype, fill in context (product, brand, audience), copy the output into your ESP.
- REST API:
POST /api/v1/generatewith output kind email_body / email_abandon_cart / email_welcome_series. API docs. - Zapier / Make: trigger on Shopify abandoned checkout, Klaviyo new profile, Airtable row - scrb returns the email copy, which Zapier can route straight back into Klaviyo as a draft.
- Bulk CSV: upload a CSV of product names + cart totals (abandoned-cart variants) and scrb generates one personalized email per row.
Pricing
- Free: 5 emails / mo.
- Starter ($9.99/mo): 100 emails = one full DTC monthly cycle (welcome series + 2 broadcasts + abandoned cart flows).
- Pro ($24.99/mo): 500 emails + bulk CSV + REST API = agencies supporting 3-5 DTC clients.
- Business ($49.99/mo): 3,000 emails = enterprise lifecycle teams managing 10+ clients or high-cadence broadcasts.