Short-lived by default.
Tokens measured in minutes, refreshed quietly. A leaked credential is a closing window, not a standing grant.
Why the gate holds
Identity, kept honest.
Everything else on the platform can fail gracefully; identity cannot. So Weke ID is built on the assumption that every input is hostile, every token will leak eventually, and every shortcut will be found.
-
01
Prove it's you
Credentials go to this origin and nowhere else — the sign-in page's policy forbids inline code, third-party hosts, and any form action that is not this server.
-
02
Mint the smallest credential
Fifteen minutes of session, and an audience grant only when a console asks for one. There is no long-lived, all-doors token to steal.
-
03
Let every service verify
The public keys are published; verification is local to each service. Weke ID cannot silently vouch for anyone — the signature either checks or it doesn't.
- Verified sign-ins
- Policy baseline
The password never travels.
It is posted once, to this origin, over TLS — never embedded in a URL, never forwarded to another host, never seen by the console you are entering.
Every door looks the same.
Sign-in, reset, and verify share one visual identity across all twelve surfaces. A door that looks unfamiliar is the phishing tell — so no legitimate door ever does.
At the door
Even forgot-password holds the line
The reset page answers neutrally whether the address exists or not, and the emailed link is single-use with an expiry — recovery is a door, so it gets the same walls.
