amina is 3Mega.ai's contributor-facing recording platform. Native Bantu-speaking communities — across Zambia, Tanzania, DRC, Malawi, Zimbabwe, Kenya, and the wider family — contribute paid, consented audio at the source. That audio becomes the substrate BantuNomics licenses to frontier AI labs. Same operating entity, same database, same schema.
amina is where the Bantu speakers who carry the languages contribute the data that makes Bantu-capable AI possible. The platform handles speaker onboarding, language and dialect declaration, demographic capture, consent recording (audit-grade, scope-granular), session management, and compensation. Contributors are paid for their work; consent is non-overridable; every recording carries a provenance chain back to its speaker and the agreement they signed.
The audio captured on amina is the audio BantuNomics licenses to frontier AI labs. The arrangement runs both ways: subscriber revenue funds the next cycle of recording, and contributors receive direct compensation that wouldn't otherwise exist for low-resource Bantu corpora.
amina captures consented audio at the source. BantuNomics licenses the resulting substrate to AI labs. Both share one operating entity, one schema, one consent standard. Subscriptions fund recording. Recordings deepen the substrate. The flywheel turns at the speed the apparatus can sustainably turn it.
The architectural separation is deliberate. amina serves contributors. BantuNomics serves subscribers. The same data flows through both, but each side has its own product, its own audience, and its own surface — so neither set of users has to think about the other to get their work done.
amina is named for the village proverb, and the village proverb is named for the work — it takes a village to raise a child, and it takes a global village to raise an AI that actually works for the whole world. Munyambala's essay sets out the framing.
amina is where the recordings happen. BantuNomics is where AI labs subscribe to the resulting substrate at $1.75M/yr — the Full Syllable Inventories, the audio bundle, the standards. Different audiences, one apparatus.