XTTS Ugandan Languages Model Architecture/Demo

Natural Speech AI

We create natural-sounding speech that understands and reflects African accents, making digital voices feel more local and human. Our primary systems support multiple Ugandan languages including Luganda, Acholi, Lunyankole, Lumasaba, and Lusoga, built on openly shared voice recordings.

Voice Generation Authentic Accents Open Access
View Model on HuggingFace
Gemma 3N DPO Evaluation / Chat Interface

Local AI Assistants

Our intelligent chat assistants are carefully trained to think and respond thoughtfully in local languages. This allows them to naturally understand context, idioms, and cultural nuances unique to Uganda, helping users intuitively solve problems.

AI Chatbots Cultural Nuance Problem Solving
Explore Gemma Model
Dataset distribution charts / VoTexUg Dashboard

Open Information Resources

Information is the foundation for African AI. We release massive, clean collections of text and voice recordings to accelerate local technology development globally and empower communities to build their own tools.

Open Data Community Resources Inclusive Tech
Access Datasets

Shared Learnings



Our work explores how to build AI that can write in African Languages and speak with authentic local accents, shared openly to advance the field.

Knowledge Sharing Community Verified Global Impact
Read Publications