AfricanVoices
DSN's AfricanVoices – a Gates-funded platform democratizing African speech data, with 1,900 hours of curated audio across Hausa, Igbo, Nigerian Pidgin, and Yoruba for voice-AI research.
AfricanVoices is DSN's platform for democratizing African speech data for research and innovation. It bridges the data gap for underrepresented African languages in AI by providing curated, richly-labelled speech datasets that researchers and developers can search, filter, and download. It is live at africanvoices.io, managed by Data Science Nigeria and supported by the Gates Foundation, and it is the direct data foundation for the Hub's Voice & Local-Language Datasets capability.
Mission
To advance inclusive AI by making high-quality African speech data accessible – closing the gap that leaves most Nigerian and African languages under-served by mainstream voice technology.
Scale
Languages covered: Hausa, Igbo, Nigerian Pidgin, and Yoruba.
What it provides
- Curated multilingual speech datasets with metadata (age, gender, domain).
- Advanced search & demographic filtering – find audio by language and speaker attributes.
- Customizable downloads – take a chunk of a dataset (5–100%) rather than all-or-nothing.
- Community feedback loop for continuous quality improvement.
Partners
Beyond DSN and the Gates Foundation, AfricanVoices is delivered with:
- Awarri Technologies
- Nigerian Institute of Translators and Interpreters
- University of Ibadan – Department of Linguistics
- RobotsMali
Who it's for
Researchers, developers, and students building voice-AI applications for African languages – ASR, TTS, and multilingual conversational systems.
Relevance to NAISH
AfricanVoices supplies exactly the data several NAISH use cases depend on. The Agricultural Advisory chatbot (Hausa / Yoruba / Igbo) and the Oral Reading Fluency assessment (local-language ASR) both need large, curated Nigerian-language speech corpora – precisely what AfricanVoices curates. It is the concrete dataset layer behind the Hub's voice & local-language capability, and a reason data readiness for voice use cases is further along than it would otherwise be.
Managed by Data Science Nigeria (DSN), supported by the Gates Foundation.
DiCert MCP
DSN's MCP server that sits as an AI-DPI middleware layer over a DPI platform, exposing its data as secure, callable tools so AI workflows and voice agents can interact with DPI safely.
Dimagi
Dimagi – the social enterprise behind CommCare and Open Chat Studio, a long-standing Gates Foundation partner whose open-source, offline-first, LLM-enabled platforms for frontline health and development work map directly onto several NAISH use cases.