I Built an AI Email Tool That Checks for Spam (Because Most AI Emails Are Terrible)

I was using ChatGPT to write cold emails for my consulting work. They looked great but got zero responses.

Turns out they were scoring 35/100 on spam filters. Basically guaranteed to hit spam folders.

So I built ColdCopy - an AI email generator that checks deliverability before you send.

The Problem

I tested popular AI email tools:

• ChatGPT emails: 30-40/100 spam score
• Other AI tools: 25-45/100 spam score

They optimize for what sounds good to humans, not what passes spam filters.

The Solution

ColdCopy generates emails then runs them through SpamAssassin to catch issues:

const checkDeliverability = async (subject, body) => {
 const response = await fetch('https://spamcheck.postmarkapp.com/filter', {
   method: 'POST',
   headers: { 'Content-Type': 'application/json' },
   body: JSON.stringify({
     email: `Subject: ${subject}\n\n${body}`,
     options: "long"
   })
 });

 const result = await response.json();
 return Math.max(0, 100 - (result.score * 6));
};

If issues are found, AI automatically fixes them while preserving the core message.
Results

ColdCopy emails: 80-95/100 spam score
Other AI tools: 25-45/100 spam score

Users report 23% average improvement in deliverability scores.
Tech Stack

Next.js + Tailwind CSS
OpenAI GPT-4 for generation
Postmark's SpamAssassin API
Vercel hosting

What I Learned
Deliverability is a huge overlooked differentiator. Most AI tools compete on writing quality, but if emails don't reach inboxes, quality doesn't matter.

Try it: coldcopy.xyz
What boring-but-critical problems are you solving?