A visitor gets one code, and that code is written to every lock the unit has — front door, gate, mailbox. It starts when the tour window opens and stops when the window closes. Handing out three codes for three locks is how a self-guided tour turns into a phone call.
Why is a unit more than one door?
Because almost none of them are. A garden building has a gate before it has a front door. A mailbox has its own lock, and a walkthrough that cannot open it is a walkthrough that ends with a question. Add a package room or a side entrance and a single “the code is 4826” text stops being enough.
The failure is not dramatic. The visitor stands at the gate at 8pm with a code for the apartment door, and calls the number on the listing. Somebody answers, or nobody does.
What does one code actually mean?
It means the code is a property of the tour, not of a lock. When the tour is booked and the visitor’s identity checks out, the same value is written to every lock attached to that unit — currently TTLock and Sifely hardware, plus lockboxes. The visitor learns one number, uses it wherever they meet a lock, and never has to know which brand is on which door.
It also means revocation is one action. A cancelled tour removes one code from every lock at once, rather than leaving whichever one somebody forgot.
What happens when a lock is offline?
Smart locks lose their connection. A basement gate with no signal is normal, not exceptional, and an access system that assumes a live connection fails exactly where the building is hardest to reach.
Offline locks get a daily fallback code instead. It is narrower than a live code — a day rather than a tour window — but it is a code that was computed rather than fetched, so it works with nothing listening. A unit with a dead spot is still tourable.
Why does the window close on its own?
Because an access code that has to be removed by hand is an access code that stays. The end of the tour window is the end of the code, and nothing has to remember to clean it up.
This is the part that separates a time-boxed code from a lockbox. A lockbox combination is shared, permanent until somebody changes it, and unattributable: the person who used it last is the person who used it forever. A code that expires answers the question a lockbox cannot — who, exactly, and when.
What does the log hold?
Every code that ever touched a lock, with its timestamp, scoped to the organisation that owns the unit. Created, used, expired.
That record is the reason the whole mechanism is worth building rather than working around. An owner asking who was in unit 4B on Tuesday gets an answer from the system instead of an answer from somebody’s memory.