Security
Security posture, stated plainly.
NocturnalPost is an email-operations platform, so its security model is the product. Here is exactly how credentials, mail data, and tenant boundaries are handled.
Credentials
All provider access uses OAuth — Microsoft and Google issue tokens; users never hand passwords to NocturnalPost. Tokens are encrypted at rest and revoked server-side on disconnect (RFC 7009). IMAP credentials, where a server requires them, are encrypted and never displayed again after entry.
Tenant isolation
Every tenant gets an isolated workspace and isolated database. Cross-tenant access is denied at the middleware, router, and database layers — and specifically tested. Within a tenant, mailboxes are personal: members control only their own.
Mail handling
Metadata-first: subject, sender, timestamps, and short previews. No tenant-wide crawling, no full-body warehousing. Connections to mail servers are TLS-only with connection timeouts; access is limited to folders the user explicitly selects.
Retention as enforcement
Retention is not a policy document — it runs hourly and deletes message data, derived content, and search vectors on the schedule the tenant chooses. Disconnect revokes tokens and stops all syncing immediately.
Transparency
The whole platform is AGPL-3.0: the security model is verifiable in source, not promised in a PDF. The privacy page details exactly what is and isn't stored. Every admin action is audit-logged.
Reporting
Found something? Report it — contact details are on the contact page. Security reports are triaged before feature work, every time.