Reads the signals, not just the words
Sender behavior, payment pressure, credential requests, link mismatches, and impersonation patterns — not keyword blocklists.
Your mom and dad shouldn't be scammed.
MailSubsystem watches the inbox for the patterns that get people in trouble — payment pressure, sender impersonation, suspicious links — and holds the risky messages for a human to look at before anything moves. We run the service so the people you care about don't have to.
Scams imitate people, exploit urgency, and borrow brand language. Spam filters weren't designed to catch them. We were.
Sender behavior, payment pressure, credential requests, link mismatches, and impersonation patterns — not keyword blocklists.
Anything risky is held in a clear review state before it moves. Nothing important gets quietly filed away.
Normal threads with known senders stay out of the way. The unusual asks are the ones that get flagged.
The intelligence comes from the open-source Core, so anyone can audit how it works. We take care of running it: account isolation, audit trails, secret handling, and the operational care a service needs when other people's mail flows through it.
Every account is its own boundary. No shared state across tenants, no cross-account access without an explicit grant that's logged.
A family member can see the risk feed and review queue without having the full inbox handed to them. Visibility scoped, not shared.
Mailbox credentials are reference-held, never written to logs, and run on Cloudflare's edge platform with strict isolation between tenants.
Join the early-interest list for product updates, beta timing, and the first chance to put MailSubsystem in front of the people you're trying to protect.