Malayalam Voice Assistant with Ollama

A voice-enabled AI assistant with native Malayalam support, powered by Ollama and local language models. This project enables natural voice interactions in both English and Malayalam, with automatic translation between languages.

🌟 Key Features

• Bilingual Support: Seamlessly understand and respond in both Malayalam and English
• Voice Interface: Speech-to-text and text-to-speech with real-time translation
• Local Processing: Runs entirely on your machine with Ollama, ensuring data privacy
• Lightweight: Uses efficient models that run well on consumer hardware
• Easy to Use: Simple command-line interface for quick interactions

🚀 Quick Start

Prerequisites

• Python 3.8 or higher
• Ollama installed and running locally (download from ollama.ai)
• A working microphone and speakers

Installation

1. Clone this repository:

   git clone https://github.com/yourusername/malayalam-ai-saas.git
   cd malayalam-ai-saas
   

2. Create and activate a virtual environment:

   python -m venv venv
   .\venv\Scripts\activate  # On Windows
   # OR
   source venv/bin/activate  # On macOS/Linux
   

3. Install the required dependencies:

   pip install -r requirements.txt
   

4. Pull the required Ollama model:

   ollama pull mistral:7b-instruct
   

Usage

1. Start the voice assistant:

   python voice_assistant_nllb.py
   

2. Press Enter to start recording your voice

3. Speak clearly in either English or Malayalam

4. The assistant will process your request and respond accordingly

🏗️ Project Structure

malayalam-ai-saas/
├── voice_assistant_nllb.py  # Main voice assistant script
├── requirements.txt         # Python dependencies
├── recordings/             # Directory for saved audio recordings
└── README.md               # This documentation file

🚀 Getting Started

Prerequisites

• Python 3.10+
• Ollama installed and running
• FFmpeg (for audio processing)

Installation

1. Clone the repository and set up the environment:

   git clone https://github.com/your-username/malayalam-ai-ollama.git
   cd malayalam-ai-ollama
   python -m venv venv
   .\venv\Scripts\activate  # On Windows
   source venv/bin/activate  # On macOS/Linux
   

2. Install the required Python packages:

   pip install -r requirements.txt
   

3. Download the Ollama model (if not already downloaded):

   ollama pull mistral  # Or any other model you prefer
   

4. Start the FastAPI server:

   uvicorn main:app --reload
   

5. Open your browser and go to http://localhost:8000/docs to test the API endpoints.

🛠️ Usage

Testing the API

1. Start the server if not already running:

   uvicorn main:app --reload
   

2. Use the interactive API documentation at http://localhost:8000/docs to test the endpoints.

Testing Ollama Integration

Run the test script to verify everything is working:

python test_ollama.py

Cleaning Up

To clean up old files and set up the new project structure:

python cleanup_and_setup.py

2. Install dependencies:

   pip install -r requirements_minimal.txt
   

3. Set up environment variables:

   cp .env.example .env
   # Update the .env file with your configuration
   

🎯 Core Features

1. Malayalam Speech Recognition

• High-accuracy Automatic Speech Recognition (ASR) for Malayalam
• Real-time speech-to-text conversion
• Noise reduction and accent adaptation

2. Intelligent Response Generation

• Context-aware responses using RAG technology
• Business-specific knowledge base integration
• Multi-turn conversation handling

3. Text-to-Speech

• Natural-sounding Malayalam speech synthesis
• Expressive and clear voice output
• Adjustable speaking rate and tone

4. Business Dashboard

• Conversation analytics and insights
• Performance metrics and KPIs
• Easy knowledge base management

🛠️ Development Roadmap

Phase 1: Core Platform (Current)

• Basic Malayalam text processing
• Translation pipeline
• Malayalam ASR integration
• Basic RAG implementation
• API endpoints for core functionality

Phase 2: Business Features

• Multi-tenant architecture
• Business onboarding workflow
• Knowledge base management
• Basic analytics dashboard
• Billing and subscription management
• Role-based access control

Phase 3: Advanced Features

• Real-time call monitoring
• Sentiment analysis
• Automated quality assurance
• Custom voice models
• Integration with popular CRMs

Phase 4: Scale & Optimization

• Auto-scaling infrastructure
• Advanced analytics and reporting
• Custom ML model training
• Multi-region deployment

🚀 Getting Started

Prerequisites

• Python 3.10+
• FFmpeg
• PortAudio (for audio processing)

Installation

1. Clone the repository:

   git clone https://github.com/yourusername/malayalam-ai-saas.git
   cd malayalam-ai-saas
   

2. Set up a virtual environment:

   python -m venv venv
   source venv/bin/activate  # On Windows: .\venv\Scripts\activate
   

3. Install dependencies:

   pip install -r requirements.txt
   

4. Set up environment variables:

   cp .env.example .env
   # Edit .env with your configuration
   

Running the Application

Start the development server:

uvicorn backend.main:app --reload

📚 Documentation

For detailed documentation, please visit our Documentation Portal.

🤝 Contributing

We welcome contributions! Please see our Contributing Guidelines for details.

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🌐 Connect With Us

• Website
• Twitter
• LinkedIn
• Blog

🙏 Acknowledgments

• Built with ❤️ in Kerala, India
• Special thanks to the open-source community for their invaluable contributions

Phase 3: Advanced Capabilities

• Sentiment analysis
• Intent recognition
• Payment integration
• Advanced analytics

🤝 Contributing

We welcome contributions from the community! Please see our Contributing Guidelines for more details.

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

📞 Contact

For business inquiries or support, please contact us at contact@malayalam-ai.com

License

This project is licensed under the MIT License - see the LICENSE file for details.