abdulsahabi/HausaVoice

Open-source Hausa STT & TTS models for voice AI, built by and for our people # HausaVoice 🔊🇳🇬 **HausaVoice** is an open-source **voice AI library** focused on building accurate and accessible **Hausa Speech-to-Text (STT)** and **Text-to-Speech (TTS)** models. Built by Hausa developers, for Hausa speakers. --- ## 💡 Mission To empower African developers and researchers with powerful, flexible, and accurate Hausa voice tools — enabling them to build smart apps, voice assistants, and learning tools in Hausa. We are committed to closing the language gap in AI by making Hausa a first-class language in speech technology. --- ## 🔧 What HausaVoice Is - A developer-friendly **Python package / Node.js module** - Provides APIs for: - 🎙️ **Hausa Speech-to-Text (STT)** - 🔊 **Hausa Text-to-Speech (TTS)** - Built with community-curated datasets, fine-tuned models, and native Hausa audio - Easily integratable in **voice apps**, **bots**, **assistants**, and **educational tools** --- ## ✅ Features (Planned) - [x] Whisper-based Hausa transcription (baseline) - [x] gTTS Hausa voice output (TTS) - [ ] Convert audio from `.aac`/`.mp3` to `.wav` - [ ] Build CLI & web demo - [ ] Custom Hausa speech datasets (open-source) - [ ] Fully-trained native Hausa STT & TTS models - [ ] Deploy as installable package (via pip/NPM) - [ ] REST API + WebSocket support --- ## 📦 Dependencies - Faster-Whisper - gTTS - ffmpeg (for audio format conversion)