We build the translation APIs, NLP tools, and language models that enable organizations to operate in African languages at scale.
To build verified AI infrastructure for every African language — starting with the most underserved. We believe the next billion internet users will speak languages that AI doesn't understand yet. MEYNG is building the bridge.
A future where every African language has dependable, verified NLP tools — translation APIs, speech recognition, and language models — accessible to any developer, organization, or government that needs them.
Africa has 2,000+ languages spoken by 1.4 billion people, and only a small fraction have basic NLP tools. Free translation coverage has grown — Google added Sango in 2026, and that is genuinely good news. But coverage is not the same as usability: organizations working in these languages still cannot verify what machines generate. That verification layer is where MEYNG operates.
2,000+
Languages
1.4B
Speakers
<5%
Have NLP
5M+
Sango speakers
Community-verified vocabulary database, published vocabulary-augmented prompting methodology, and an open verification pipeline. The moat is data quality and cultural authority — not just compute.
Founded by a Central African native speaker with deep linguistic expertise. Our data quality comes from cultural understanding, not just computation.
We publish our data with per-entry provenance and our evaluations with the failures included — building the verification standard any Sango AI, ours included, gets measured against.
Built on world-class foundation models, published openly for AI community visibility, deployed on the world's largest messaging platform for maximum reach across Africa.
Founder & CEO
Software engineer and AI entrepreneur. Native Sango speaker from the Central African Republic. Built SangoAI from zero to a live platform — published NLP toolkit, open verified datasets, and vocabulary-augmented prompting research. Invited by BPI France to Africa Forward Summit, Nairobi 2026. Based in Paris.
Incorporated in Paris. Began building a Sango NLP pipeline and training data collection.
Translation, chat, and learning platform launched at sangoai.sbs.
Open-source NLP toolkit published for developers. Vocabulary-augmented prompting methodology and verified Sango lexicon released openly. Sango becomes a documented NLP research language.
eNdara learning bot published on Meta. Any WhatsApp user can message to learn Sango. Real-world deployment proof beyond the web.
We're raising pre-seed funding to scale from 1 language to 5, hire our first NLP engineers, and close our first pilot contracts. The need is vast, the timing is now.
Let's Talk