Today we're announcing that we're sunsetting new.email.
The product will remain available for you to export your templates until August 7, 2026. After that, new.email will no longer accept new prompts or sends. If you need help migrating, you can reach out to us.
I wanted to take a moment to talk about why we're doing this, what we would do differently next time, and what we would do exactly the same.
Why we built it
Every day, product teams need to send emails to their users. Some teams have a developer to build their templates. Many depend on marketing, product, or design. The result is often a mix of emails that don't look or feel like the rest of the product.
We believed emails should be consistent, beautiful, and responsive, regardless of who builds them.
In February 2025, we announced new.email, a tool where anyone could create email templates using natural language, powered by React Email components and the deliverability lessons we learned from sending millions of emails every day.
At that time, Claude Code and Codex didn't exist. People were using GPT-4 and Sonnet 3.5, and those models weren't as good as they are today for email creation.
The core bet was right. But we also made two decisions early on that made sustaining the product harder than it needed to be.
What we would do differently
Lesson 1: New brands require a ton of investment
We launched new.email with a completely separate identity. New name, new domain, new logo, new visual language.
This was intentional. We thought we were targeting a different audience than Resend, so a different brand made sense on paper.
In practice, maintaining one high-quality brand is already a lot of work, especially with a small team. Every brand needs its own website, its own docs, its own social presence, its own design decisions, and its own care.
In retrospect, we would fold everything into the Resend brand from day one. A different audience doesn't necessarily require a completely different identity.
Lesson 2: A separate stack is fast to launch and slow to live with.