Skip to content

← Notes

Invalid Date

There are already a handful of email-testing services out there, and they’re fine. So why another one?

Three reasons:

  1. Most teams still hand-roll some version of this. Either they pay for a heavyweight QA email suite they barely use, or they stand up an inbox and re-write the same polling helper they wrote at the last job. The helper should come pre-written.

  2. Agent evaluation needs a primitive that didn’t exist when the incumbents were built. Eval wants throwaway inboxes that can be re-seeded deterministically — something the older tools are retrofitting, not designing for. MailFade was scoped around it from day one.

  3. The whole thing fits on the edge. Cloudflare Email Routing, an Email Worker, D1 and R2, and an API Worker in front. No SMTP server. No queue. No dedicated host. The marginal cost of an extra message rounds to zero.

In v1

  • A catch-all on @mailfade.dev. Every address is live the instant the first message arrives.
  • A five-endpoint HTTP API: list, message, message-raw, attachment, keys.
  • A real free tier — 100 requests per day, no key required — for wiring it into side projects.
  • Paid plans for serious CI use, with custom-domain support on Team and up.
  • Card subscriptions over Stripe and one-time top-ups over Lightning, with a single shared key.

What’s next

  • A handful of framework-specific helpers published as actual npm and PyPI packages, so the recipe stops being copy-paste from the docs.
  • Webhook callbacks on delivery, so an agent can react in milliseconds instead of polling.
  • A small CLI wrapper for local development.

If you build something with it, or have a use case worth designing for, write to support@mailfade.dev.