kukin01/Kinyarwanda_whisper_voice_AI

# Kinyarwanda_whisper_voice_AI A complete voice interaction system for Kinyarwanda, simulating humanoid robot communication (ASR → NLP → TTS pipeline). ## Features - 🎤 **Speech Recognition**: Convert Kinyarwanda speech to text - 🧠 **Question Understanding**: Match queries to answers - 🔊 **Voice Responses**: Natural-sounding Kinyarwanda output - 🌐 **Web Interface**: Easy-to-use Gradio UI ## Prerequisites - Python 3.8+ - Microphone (for live recording) - FFmpeg (for audio processing) ## Installation 1. Clone the repository: ```bash git clone github.com cd kinyarwanda-voice- kinyarwanda-voice-assistant ├── config/ │ ├── intents.json # Conversation patterns │ └── qa_pairs.json # Question-Answer pairs ├── asr_module/ │ ├── audio_samples/ # Input recordings (5+ .wav files) │ └── responses/ # Generated TTS outputs ├── modules/ │ ├── asr.py # Speech recognition │ ├── nlp.py # Language processing │ └── tts.py # Speech synthesis ├── main.py # Main application └── requirements.txt # Dependencies Audio Samples Included samples in /data/audio_samples: amakuru.wav - "Amakuru?" witwa_nde.wav - "Witwa nde?" uri_he.wav - "Uri he?" ushobora_gukora_iki.wav - "Ushobora gukora iki?" umunsi_mwiza.wav - "Umunsi mwiza?" Recording Guidelines: Sample rate: 16kHz Format: 16-bit PCM WAV Duration: 2-5 seconds per phrase 🚀 Usage Running the Assistant bash python main.py Interacting via Gradio UI Access the web interface at localhost after launching: Click the microphone to record Or upload pre-recorded audio System will respond with voice and text Testing Components Individually ASR Test: python from modules.asr import Transcriber transcriber = Transcriber("data/audio_samples/amakuru.wav") print(transcriber.transcription) TTS Test: python from modules.tts import kinya_tts kinya_tts("Muraho!")