Posthaste

Security

Your mail, protected in ways you can check.

Specific claims, how each one works, and the things we don’t claim. Written by the people who built it.

Isolation the database enforces.

Every customer table is fenced by Postgres row-level security, with FORCE on, and the app connects as a role that cannot bypass it. One account’s query can’t return another account’s rows, even if our own code asks for them.

Bodies encrypted at rest.

AES-256-GCM, a fresh nonce per value, the key kept outside the database. Subjects, addresses and headers aren’t encrypted, and we say so.

We can’t read your mail without telling you.

Opening a body takes a written reason, writes an audit row in the same transaction, and emails your account owner the reason word for word. It can’t be switched off, not even by us.

A record nobody can quietly edit.

Each delivery event carries the hash of the one before it, so a missing or changed entry shows. Our billing ledger and staff audit log work the same way.

  1. #1479 acceptedprev 9a41…hash 07fe5c3b…
  2. #1480 queuedprev 07fe…hash 42d8aa19…
  3. #1481 attemptedprev 42d8…hash e1b3f6d2…
  4. #1482 deliveredprev e1b3…hash 9c41e07a…

API keys are never stored.

We keep an HMAC of each key under a server-side pepper. You see a key once, when it’s made, and we can’t recover it.

Signing keys, sealed.

A 2048-bit DKIM key per domain, encrypted at rest with AES-256-GCM.

Signed webhooks.

Every event carries a timestamped HMAC signature over the raw body, so your app can reject forgeries and replays.

Encrypted on the way.

The API only speaks HTTPS. Mail to inbox providers goes over TLS whenever the receiving server offers it.

Backed up, and restored for practice.

Nightly backups kept for fourteen days, encrypted off-site, with restores we actually rehearse. Hosted in the EU, operated from India.

Sign-in that resists a stolen password.

Two-factor with an authenticator app, passkeys, or Google sign-in. Every staff account must use two-factor.

See who processes data for us

Plainly

What we don’t claim.

Security pages often imply more than is true. Here’s what isn’t in place yet, so you can decide with the full picture.

  • No SOC 2 or ISO 27001 certification.
  • No independent security audit yet.
  • No paid bug bounty. We do thank, credit and protect people who report in good faith.

Found a vulnerability?

Write to [email protected]. We acknowledge every report within three working days, and we won’t pursue anyone who tests in good faith and tells us first.

Email security
Security disclosures
[email protected]
Abuse reports
[email protected]
Data protection
[email protected]

Bring your security questionnaire.

We’ll answer it straight, including the parts we haven’t done.