Intron, a Nigerian AI startup founded in 2020 by Tobi Olatunji and Olakunle Asekun, expanded its Sahara speech-recognition platform to 57 languages covering more than 500 distinct African accents, with the company claiming up to 64% better accuracy than Gemini, GPT-4, Whisper, ElevenLabs, and Azure on African names, organisations, and locations in its own benchmarks. "We are a for-profit company, so we prioritised languages where there is enterprise demand and willingness to pay," Olatunji, founder and CEO of Intron, told TechCabal.
Sahara v2 adds 24 languages to the prior lineup, including Hausa, Swahili, isiZulu, Yoruba, Kinyarwanda, Twi, Igbo, isiXhosa, African French, Amharic, Bemba, Luganda, Oromo, Pidgin, Shona, and Wolof. Twenty-three of the 57 languages are African. Independent benchmarks commissioned with African voice datasets show 35% better performance with numbers, 20% more robustness in noisy and multi-speaker environments, and 25% higher cross-domain accuracy across healthcare, finance, legal services, and telecommunications.
Intron began with clinical documentation tools before moving into broader voice infrastructure; today it reports deployment in Nigeria, South Africa, Ghana, Uganda, Rwanda, and Kenya. Enterprise customers include the Ogun State Judiciary in Nigeria and Audere, a South African company that uses Sahara to transcribe WhatsApp voice messages across local accents. The company also shipped what it describes as the world's first bilingual Swahili-English automatic speech recognition model, built in partnership with Kenyan outpatient provider Penda Health to handle real-time code-switching during clinical conversations. A first local-language text-to-speech model in Hausa followed.
Sahara v2 was trained on more than 14 million audio clips totaling over 50,000 hours from 40,000+ speakers across 30 African countries. Cloud is the primary deployment mode, with offline enterprise deployments supported through a partnership with Nvidia on Jetson Edge devices — the entry-level unit costs about $250 and serves several users connected locally. Intron plans to raise $3 million later in 2026 to expand language coverage and continue building bilingual and domain-specific models in Yoruba, Hausa, Zulu, and Kinyarwanda. The company also publishes the inaugural 2026 Africa Voice AI Report, which examines benchmarking standards and commercial readiness of voice systems across the continent.