Every campaign I've shipped started the same way: open the template library, pick something close enough, swap in the copy, eyeball the mobile preview, send it, then find out the CTA button was the wrong shade of gray on Android. The feedback loop is one send long.
I built kopi to close that loop before export. You write a brief, it generates a complete HTML email in under 5 minutes, then runs an automated design critique across 12+ criteria — layout hierarchy, mobile readability, CTA contrast — and gives you a 0–100 score. The score isn't decorative; it surfaces specific things to fix before the campaign goes anywhere.
The angle I want to explain here is the brand system piece, because it's what makes the scores actionable rather than generic.
Most design critiques are context-free. A 60/100 on CTA contrast means the button fails WCAG minimums against some default background. But if your brand uses a navy button on a white card, the relevant threshold is different from someone running orange-on-black. kopi learns your fonts, colors, and layout preferences and applies them when it scores. So a contrast flag means your specific button color failed against your specific background — not a hypothetical.
The same system drives generation. When you write a brief, kopi doesn't start from a generic template and ask you to re-skin it. It generates inside your design system from the first pass. The 400+ emails in the public gallery at trykopi.ai/emails that score 80+ out of 100 are generated that way — brand-aware from prompt, not patched after the fact.
Klaviyo export is direct. If you work in Claude Code or Cursor, there's an MCP server.
Top comments (0)