Wa-Zo-Bia, the voice-consent registry

Field notes

The proposed workflow combines browser-based speaker signatures with a review step in which a named person listens to the clip and checks its transcript. Records retain scope, expiry, and three time representations, while the page says withdrawals should propagate across the system. The language-handling detail is unusually concrete: text remains UTF-8 and is not normalized, preserving tone marks and underdots in exports. Registry counts are not yet populated and appear as dashes.