🎙️ AI Voice Studio

Multilingual Speech AI for TTS, ASR, and practical audio processing.

AI Voice Studio supports Uzbek, English, and Korean and demonstrates open-source speech-model integration, resource-aware inference, multilingual UI, and Hugging Face deployment.

Live demo: https://huggingface.co/spaces/IKROMJON01/AI-voice-Studio

• Uzbek, English, Korean
• Hugging Face MMS TTS models
• Lazy model loading

• Whisper Large V3 Turbo
• Uzbek / English / Korean
• Upload or microphone input

• Noise reduction
• Silence trimming
• Peak normalization
• WAV export

The public interface contains the cloning architecture, while the cloning model remains disabled until a suitable multilingual model is integrated and evaluated.

Only use voice cloning with the speaker's permission.

Models are loaded lazily to reduce unnecessary CPU/RAM/VRAM use.

AI Voice Studio
       |
 +-----+---------+
 |       |       |
 v       v       v
TTS     ASR    Audio Processing
 |       |       |
 +-------+-------+
         |
         v
    Gradio App
         |
         v
 Hugging Face Spaces

Python · PyTorch · Transformers · Gradio · Hugging Face Spaces · Whisper · Facebook MMS TTS · Librosa · SoundFile · SciPy

• Lazy loading for large speech models
• CPU/GPU-compatible inference paths
• Multilingual model routing
• Lightweight audio preprocessing
• Hugging Face deployment
• Interactive Gradio UI

git clone https://github.com/ikromjon-gif/Ai-voice-studio.git
cd Ai-voice-studio
pip install -r requirements.txt
python app.py

• Use voice cloning only with your own voice or explicit permission.
• Do not use generated audio for unauthorized impersonation.
• Treat synthetic speech as generated content.

Speech AI · TTS / ASR · Transformers · Multilingual AI · Audio Processing · Resource-aware inference · Model integration · Hugging Face deployment

Author: Ikromjon Tojiboev