Our own identity layer

Detalles

The client is not named here; the architecture is the story.

Signing in to one product means signing in to all six. Nobody types a password: you identify yourself by email or phone, confirm with a one-time code, then open the session with a security code you chose. No password is ever stored, so there is no password to leak.

The system returns identity only — who you are, your name, email, phone. The question "is this user an admin?" is never answered centrally; each application resolves authorization in its own database, keyed by the stable user id the provider hands back. That separation is deliberate: when one product changes its permission model, the identity layer is untouched.

Every application gets its own consent screen, and users can revoke a grant from their own profile. Sessions live on real activity — a blind heartbeat cannot keep one alive forever. When an administrator suspends an account or a user signs out, every connected application learns about it on the same round trip.

Device tracking, login history and per-identifier pauses make it visible where an account is being opened from. Auth codes are single-use, live thirty seconds and are bound to the address they were issued for, so a leaked code cannot be exchanged anywhere else.