Audio Studio AI 🎤

Audio Studio AI is an interactive application built with Python and Streamlit, allowing users to generate high-quality audio content locally using advanced text-to-speech technology. It features both a beautiful web interface and a powerful REST API for integration. Perfect for creating voiceovers for videos, podcasts, and other audio content.

🚀 Features

• Local AI-powered text-to-speech with high-quality voice synthesis
• Multiple language support with various voices for each language
• Sentence-based audio generation with customizable pauses
• Individual sentence preview and editing
• Export/Import functionality for sentence configurations
• Multiple output formats (WAV, MP3) with dynamic format support
• Customizable speech speed for each sentence
• Beautiful Streamlit web interface for easy interaction
• REST API for integration with other applications
• Local processing - no cloud dependencies required

📞 Installation

1. Clone the Repository

git clone https://github.com/paulocoutinhox/audio-studio-ai.git
cd audio-studio-ai

2. Create a Virtual Environment

python3 -m venv .venv
source .venv/bin/activate  # macOS/Linux
.venv\Scripts\activate     # Windows

3. Install Dependencies

pip install -r requirements.txt

4. Download Required Model Files

Download the required model files from the model repository and place them in the models/ directory of the project. See the Model Support section for details.

⚙️ Configuration

1. Model Files

The application expects the model files to be in the models/ directory. See the Model Support section for specific file requirements.

2. Output Configuration

You can configure the following settings in the sidebar:

• Output format (WAV or MP3)
• Minimum and maximum pause duration between sentences
• Model and voices file paths

🛠️ Usage

Web Interface (Streamlit)

1. Run the Streamlit Application

   streamlit run app.py

2. Steps in the Web UI

   • Add sentences using the "Add Sentence" button
   • For each sentence:
     • Enter the text
     • Select the language
     • Choose a voice
     • Adjust the speech speed
   • Use the up/down arrows to reorder sentences
   • Delete sentences using the trash icon
   • Click "Generate Audio" to create the final audio
   • Preview individual sentences or download the complete audio

REST API

1. Start the API Server

   python api.py

2. API Documentation

   • Server runs on: http://localhost:8000
   • Interactive docs: http://localhost:8000/docs
   • 📚 Complete API Documentation - Examples, endpoints, and integration guides

3. Quick API Example

   curl -X POST "http://localhost:8000/generate-audio" \
     -H "Content-Type: application/json" \
     -d '{
       "sentences": [{
         "text": "Hello world!",
         "lang": "en-us",
         "voice": "af_sarah",
         "speed": 1.0
       }],
       "output_format": "mp3"
     }'

🌍 Supported Languages and Voices

The application supports multiple languages with various voices for each:

• American English (en-us)
  • Multiple voices including af_sarah, af_nova, af_river, and more
• British English (en-gb)
  • Voices like bf_alice, bf_emma, bm_daniel, and more
• Japanese (ja)
  • Voices including jf_alpha, jf_gongitsune, jm_kumo, and more
• Mandarin Chinese (zh)
  • Multiple voices like zf_xiaobei, zf_xiaoxiao, zm_yunjian, and more
• Spanish (es)
  • Voices including ef_dora, em_alex
• French (fr)
  • Voice ff_siwis
• Hindi (hi)
  • Voices including hf_alpha, hf_beta, hm_omega
• Italian (it)
  • Voices if_sara, im_nicola
• Brazilian Portuguese (pt-br)
  • Voices pf_dora, pm_alex, pm_santa

🤖 Model Support

This application currently supports the Kokoro TTS model for high-quality text-to-speech synthesis. To use the application:

1. Download the following files from Hugging Face - Kokoro-82M:

   • kokoro-v1.0.onnx
   • voices-v1.0.bin

2. Download the model files using these direct links:

   • kokoro-v1.0.onnx
   • voices-v1.0.bin

3. Place these files in the models/ directory of the project.

The Kokoro model provides:

• High-quality voice synthesis
• Support for multiple languages
• Various voice options for each language
• Fast local processing
• No cloud dependencies

📁 Project Structure

audio-studio-ai/
│
├── 📝 README.md               # Project documentation and guide
├── 📚 API.md                  # Complete REST API documentation
├── 🎯 app.py                  # Main Streamlit application interface
├── 🚀 api.py                  # FastAPI REST API server
├── ⚙️ config.py               # Configuration and settings management
├── 🛠️ utils.py                # Utility functions and helpers
├── 📦 requirements.txt        # Project dependencies list
│
├── 🤖 models/                 # AI model files directory
│   ├── 🧠 kokoro-v1.0.onnx    # TTS neural network model
│   └── 🗣️ voices-v1.0.bin     # Voice data and configurations
│
├── 🎵 temp/                   # Temporary audio files storage
│
└── 🎨 extras/                 # Additional resources
    └── 🖼️ images/             # Images, icons and assets

🤝 Contributing

1. Fork the repository
2. Create a new branch (git checkout -b feature-xyz)
3. Commit changes (git commit -m "Added new feature")
4. Push to the branch (git push origin feature-xyz)
5. Open a pull request

📞 Contact

For issues or contributions, open a GitHub issue or contact: 💎 paulocoutinhox@gmail.com 🔗 GitHub

📜 License

MIT

Copyright (c) 2025, Paulo Coutinho