What we do
- Voz: people read prompts and speak on topics in their own accent.
- Video: first-person video of real work, with narration out loud, across 13 trades.
- Selfies and identity: identity photos for verification and, when a project asks for it, face datasets.
- Expert text: licensed professionals receive a scenario, answer it independently with a cited source, and disagreements are resolved before delivery. We publish an open regional benchmark, Savia Carga.