Two South African AI apps, including one building AI infrastructure across more than 60 African languages, are among five winners of the Llama Impact Grant for Startups and Researchers.
Unveiled at the UN General Assembly’s Unstoppable Africa 2025 event, the awards are an initiative by Meta, in partnership with Data Science Africa. It was launched in March 2025 to enable startups, researchers, and innovators across Sub-Saharan Africa to harness Meta’s open-source large language model, Llama, to tackle some of the region’s most pressing challenges.
Each winner receives $20,000 in funding alongside technical mentorship, networking opportunities, and engagement with policy and ecosystem stakeholders to accelerate growth and maximise societal impact.
The programme builds on Meta’s broader commitment to strengthen Africa’s AI and innovation ecosystems, prioritising scalable solutions in healthcare, education, agriculture, and digital accessibility.
The Winners:
- Vambo AI (South Africa) – Co-founder and CEO Chido Dzinotyiwei is building Africa’s multilingual AI infrastructure through Vambo AI, which develops both proprietary and open-source models powering translation, transcription, generation, and search across 60+ African languages. The platform treats language as critical infrastructure, accelerating digital inclusion and innovation at scale.
- PropelMapper (South Africa) – Created by Reghardt Adriaan Pretorius and Mark Donne, PropelMapper equips agriculture advisors with AI tools that generate tailored farmer podcasts, transform debriefs into professional reports, and uses satellite imagery with Llama models to provide intervention alerts, boosting farmer productivity and food security.
- Radease (Nigeria) – Taiwo Oyewole, Co-founder and CEO, leads Radease in simplifying access to safe medicines by empowering Patent and Proprietary Medicine Vendors (PPMVs) with WhatsApp-based AI tools that improve access to trusted health information and resources in underserved communities.
- TeenApp (Uganda) – Developed at Makerere University’s AI and Data Science Center, Rahman Sanya, Lecturer and Researcher, leads TeenApp, a digital health solution that delivers accurate, youth-friendly sexual and reproductive health education and services, guided by principles of Responsible AI.
- Easy Read Africa (Rwanda) – Led by Isaac Manzi, AI and Data Innovation Advisor at GIZ and Community Manager at MbazaNLP, Easy Read Africa transforms complex documents into simplified text, visuals, and natural voice narration, making information accessible for people with cognitive and learning challenges.
“We received an incredible number of applications this year, reflecting the vibrant and growing AI ecosystem across Africa. These projects exemplify the spirit of innovation and impact that the ‘Llama Impact Grant for Startups and Researchers’ stands for. We are excited to support their journeys and look forward to seeing the positive change they will bring to their communities and beyond,” said Sherry Dzinerova, Director of AMET Public Policy, Programs, Campaigns and Product, Meta, during the announcement at GABI.
Meta’s family of Llama AI models has been downloaded more than one billion times globally. The latest version, Llama 3.3, is open source, meaning it is freely available for organisations to use, modify, and build upon. Llama models are already being applied in projects ranging from localising educational material to expanding access to medical information in low-resource communities.
The Llama Impact Grant, launched in October 2023 as a global programme, seeks to source and scale innovative use cases of Llama that address critical challenges in education, health, agriculture, and digital inclusion. Since its inception, the programme has attracted over 800 applications from 90+ countries, with past finalists including Digital Green’s Farmer Chat and Jacaranda Health’s PROMPTS, both delivering AI-enabled solutions for smallholder farmers and maternal health across Sub-Saharan Africa.