omnirag 2.0.1

India's First Multi-Language Voice RAG Framework

🚀 OmniRAG v2.0 - Multi-Language Voice RAG

India's First Multi-Language Voice RAG Framework 🇮🇳

Intelligent RAG combining Liquid + Agentic + Chain architectures with unique features:

• 🌍 Smart Multi-Language Translation - 27+ languages including Tamil, Hindi
• 🎤 Voice Input & Output - Speak questions, hear answers
• 🧠 Adaptive RAG - Automatically adjusts to user expertise level

---

🆕 What's New in v2.0?

✨ Feature 1: Smart Post-Retrieval Translation

Revolutionary architecture: Documents stay in original language, translation happens AFTER retrieval!

Why this is better:

• ✅ Better embeddings (preserve semantic meaning)
• ✅ No storage duplication
• ✅ One document → Many output languages
• ✅ 70% more efficient than traditional approaches

from omnirag import OmniRAG

# Documents in English
rag = OmniRAG(output_language="Tamil")
rag.add_documents(["AI helps solve complex problems."])

# Query in English, get Tamil answer!
result = rag.query("What is AI?")
print(result['answer'])
# Output: "செயற்கை நுண்ணறிவு சிக்கலான சிக்கல்களைத் தீர்க்க உதவுகிறது."

Supported Languages (27+):

• Indian: Tamil, Hindi, Telugu, Marathi, Gujarati, Kannada, Malayalam, Punjabi, Bengali
• European: Spanish, French, German, Italian, Portuguese, Russian, Polish, Dutch, Turkish
• Asian: Chinese, Japanese, Korean, Vietnamese, Thai, Indonesian, Malay
• Other: Arabic, English

✨ Feature 2: Voice Input & Output

First open-source RAG with built-in voice support!

# Voice output (text-to-speech)
rag = OmniRAG(enable_voice=True, output_language="Tamil")
result = rag.query("What is Python?", speak_answer=True)
# Hears answer in Tamil! 🔊

# Voice input (speech-to-text) - requires microphone
result = rag.voice_query()
# Speak your question, hear the answer!

---

🎯 What is OmniRAG?

OmniRAG is an advanced Retrieval-Augmented Generation system that combines three powerful RAG techniques:

🌊 Liquid RAG

Automatically adapts answers to user expertise level:

• Beginner: Simple explanations with examples
• Intermediate: Balanced technical content
• Expert: Deep technical details

🤖 Agentic RAG

Intelligently chooses the best information source:

• VectorDB: For local documents
• Web Search: For current information

⛓️ Chain RAG

Handles complex multi-part questions:

• Breaks down complex queries
• Answers each part separately
• Synthesizes coherent final answer

---

✨ All Features

v2.0 NEW:

• 🌍 Multi-Language Translation (27+ languages)
• 🎤 Voice Input & Output (speak & hear)
• 🔤 Full Language Names ("Tamil" not "ta")
• 🔧 UTF-8 Support (perfect Tamil/Hindi display)

v1.0 CORE:

• ✅ PDF Support - Load PDF files directly
• ✅ Multiple LLM Models - Qwen, Flan-T5, Mistral, Phi-2
• ✅ FAISS Vector DB - Fast similarity search
• ✅ Web Search - DuckDuckGo integration (free!)
• ✅ Smart User Detection - Auto expertise level detection
• ✅ Query Decomposition - Handles complex questions
• ✅ Fast Caching - 3x speedup on repeated queries
• ✅ 100% FREE - No API costs!
• ✅ Works on CPU - No GPU required

---

📦 Installation

pip install omnirag

With Voice Input (Optional)

Windows:

pip install pipwin
pipwin install pyaudio
pip install omnirag[voice-input]

Mac:

brew install portaudio
pip install omnirag[voice-input]

Linux:

sudo apt-get install portaudio19-dev
pip install omnirag[voice-input]

From Source

git clone https://github.com/Giri530/omnirag.git
cd omnirag
pip install -e .

---

🚀 Quick Start

Basic Usage

from omnirag import OmniRAG

# Initialize
rag = OmniRAG(model_name="google/flan-t5-small")

# Add documents
rag.add_documents([
    "Python is a programming language.",
    "It is used for AI and data science."
])

# Query
result = rag.query("What is Python?")
print(result['answer'])

Multi-Language Example

from omnirag import OmniRAG

# Initialize with Spanish output
rag = OmniRAG(
    model_name="google/flan-t5-small",
    output_language="Spanish"  # or "Tamil", "Hindi", etc.
)

# Add English documents
rag.add_documents([
    "AI helps solve complex problems.",
    "Machine Learning is a subset of AI."
])

# Query in English, get Spanish answer!
result = rag.query("What is AI?")
print(result['answer'])
# Output: "La IA ayuda a resolver problemas complejos."

Voice Example

from omnirag import OmniRAG

# Initialize with voice
rag = OmniRAG(
    enable_voice=True,
    output_language="Tamil"
)

rag.add_documents(["Python is great for AI."])

# Text input, voice output
result = rag.query("What is Python?", speak_answer=True)
# Hears answer in Tamil! 🔊

# Voice input, voice output (requires microphone)
result = rag.voice_query()
# Speak question, hear answer!

---

💡 Usage Examples

Load Different File Types

# PDF files
rag.load_from_file("research_paper.pdf")

# Text files
rag.load_from_file("notes.txt")

# JSON data
rag.load_from_file("data.json")

# Entire folder
rag.load_from_folder("./documents")

# With chunking for large files
rag.load_from_file("big_file.pdf", chunk_size=500)

# Direct text
rag.add_documents([
    "Python is great for ML.",
    "Qwen is a powerful language model."
])

Different Output Languages

# Default: Spanish
rag = OmniRAG(output_language="Spanish")

# Query 1: Spanish (default)
result1 = rag.query("What is AI?")

# Query 2: Override to Tamil
result2 = rag.query("What is ML?", output_language="Tamil")

# Query 3: Override to French
result3 = rag.query("What is DL?", output_language="French")

Full Language Names

# All these work!
rag = OmniRAG(output_language="Spanish")  # ✅
rag = OmniRAG(output_language="spanish")  # ✅
rag = OmniRAG(output_language="es")       # ✅

# Same for all languages
rag = OmniRAG(output_language="Tamil")    # ✅
rag = OmniRAG(output_language="Hindi")    # ✅

Complex Queries

# OmniRAG automatically breaks down and answers
result = rag.query("""
Compare Python vs Java for machine learning.
Which is better for beginners?
What are the performance differences?
""")

print(result['answer'])

Enable Web Search

rag = OmniRAG(
    model_name="google/flan-t5-small",
    enable_web_search=True  # Free DuckDuckGo search
)

# Queries about "latest" or "recent" automatically use web
result = rag.query("Latest AI developments in 2025")

---

🎨 Supported Models

Qwen Models (Recommended!)

# Fast & Efficient
rag = OmniRAG(model_name="Qwen/Qwen2.5-0.5B-Instruct")

# Balanced (Best Choice!)
rag = OmniRAG(model_name="Qwen/Qwen2.5-1.5B-Instruct")

# High Quality
rag = OmniRAG(model_name="Qwen/Qwen2.5-3B-Instruct")

Flan-T5 Models

# Small & Fast
rag = OmniRAG(model_name="google/flan-t5-small")   # 80M params

# Medium
rag = OmniRAG(model_name="google/flan-t5-base")    # 250M params

# Larger & Better
rag = OmniRAG(model_name="google/flan-t5-large")   # 780M params

---

🏗️ Architecture

User Query
    ↓
🌊 LIQUID RAG: Detect expertise level
    ↓
⛓️ CHAIN RAG: Break into sub-queries (if complex)
    ↓
FOR EACH SUB-QUERY:
    ↓
🤖 AGENTIC RAG: Choose tool (VectorDB or Web)
    ↓
    Retrieve relevant chunks (ORIGINAL language)
    ↓
🌊 LIQUID RAG: Transform to user level
    ↓
    Generate sub-answer
    ↓
⛓️ CHAIN RAG: Synthesize all sub-answers
    ↓
🌍 TRANSLATION: Convert to target language (NEW!)
    ↓
🔊 VOICE: Speak answer (if enabled) (NEW!)
    ↓
✨ Perfect Answer!

---

📊 Performance

Model	Size	RAM	Speed	Quality
flan-t5-small	80M	0.5GB	⚡⚡⚡	⭐⭐
flan-t5-base	250M	1GB	⚡⚡⚡	⭐⭐⭐
Qwen-0.5B	0.5B	1GB	⚡⚡	⭐⭐⭐
Qwen-1.5B	1.5B	2GB	⚡⚡	⭐⭐⭐⭐
Qwen-3B	3B	4GB	⚡	⭐⭐⭐⭐⭐

Recommended:

• For testing: flan-t5-small (fast!)
• For production: flan-t5-base or Qwen-0.5B (balanced)
• For quality: Qwen-1.5B (best!)

---

🔧 Configuration

rag = OmniRAG(
    # LLM Model
    model_name="google/flan-t5-small",
    
    # Embedding Model
    embedding_model="all-MiniLM-L6-v2",
    
    # Web Search
    enable_web_search=True,
    
    # NEW: Output Language
    output_language="Tamil",  # or "auto" for no translation
    
    # NEW: Voice I/O
    enable_voice=True,
    
    # Verbose Output
    verbose=True
)

---

📖 API Reference

OmniRAG Class

__init__(model_name, embedding_model, enable_web_search, verbose, output_language, enable_voice)

Initialize OmniRAG system.

New Parameters:

• output_language (str): Target language ("Tamil", "Spanish", "auto", etc.)
• enable_voice (bool): Enable voice input/output

query(user_query, output_language=None, speak_answer=False)

Query the system and get answer.

New Parameters:

• output_language (str): Override default language for this query
• speak_answer (bool): Speak the answer aloud

Returns:

{
    'answer': str,              # Generated answer
    'sources': list,            # Retrieved sources
    'user_level': str,          # Detected expertise level
    'output_language': str,     # Output language code
    'spoken': bool,             # Whether answer was spoken
}

voice_query(output_language=None) NEW!

Voice-to-voice query (requires microphone).

save_to_file(result, filename) NEW!

Save result to file with UTF-8 encoding.

Other Methods (from v1.0)

• load_from_file(file_path, chunk_size=None)
• load_from_folder(folder_path, file_extensions=None)
• add_documents(documents)
• get_stats()
• clear_cache()

---

🌍 Use Cases

Customer Support (Multi-Language)

rag = OmniRAG(output_language="Hindi", enable_voice=True)
rag.load_from_file("product_manual.pdf")

# Hindi-speaking customer
result = rag.query("How do I reset my device?", speak_answer=True)
# Answer in Hindi + spoken aloud!

Educational Platform (Tamil)

rag = OmniRAG(output_language="Tamil")
rag.load_from_file("class10_science.pdf")

# Student query
result = rag.query("What is photosynthesis?")
# Answer in Tamil!

Accessibility Tool

# For visually impaired users
rag = OmniRAG(enable_voice=True)
rag.load_from_folder("./personal_docs")

# Completely hands-free
while True:
    result = rag.voice_query()
    if "exit" in result.get('answer', '').lower():
        break

---

🌟 Why OmniRAG?

Feature	LangChain	LlamaIndex	OmniRAG
Post-Retrieval Translation	❌ No	❌ No	✅ YES
Built-in Voice I/O	❌ No	❌ No	✅ YES
Indian Language Support	⚠️ Basic	⚠️ Basic	✅ Native
Full Language Names	❌ No	❌ No	✅ YES
Beginner Friendly	⚠️ Complex	⚠️ Complex	✅ Simple
100% Free	✅ Yes	✅ Yes	✅ Yes

---

🛠️ Development

Install for Development

git clone https://github.com/Giri530/omnirag.git
cd omnirag
pip install -e ".[dev]"

Project Structure

omnirag/
├── omnirag/
│   ├── __init__.py
│   ├── omnirag.py              # Main class
│   ├── smart_translator.py     # NEW: Translation
│   ├── voice_processor.py      # NEW: Voice I/O
│   ├── liquid_analyzer.py      # User level detection
│   ├── chain_decomposer.py     # Query decomposition
│   ├── agentic_planner.py      # Tool selection
│   ├── content_transformer.py  # Content adaptation
│   ├── vectordb_tool.py        # FAISS database
│   ├── web_search_tool.py      # Web search
│   ├── llm_client.py           # LLM wrapper
│   └── cache.py                # Caching
├── examples/
│   └── quickstart.py
├── setup.py
├── pyproject.toml
├── requirements.txt
└── README.md

---

📝 Requirements

• Python 3.8+
• 1-4GB RAM (depends on model)
• CPU or GPU (GPU recommended for speed)

Core Dependencies:

• transformers, torch, sentence-transformers
• faiss-cpu, PyPDF2, duckduckgo-search

New Dependencies (v2.0):

• deep-translator, langdetect (translation)
• pyttsx3 (voice output)
• SpeechRecognition, pyaudio (voice input - optional)

---

🤝 Contributing

Contributions welcome! Please:

1. Fork the repository
2. Create feature branch (git checkout -b feature/amazing)
3. Commit changes (git commit -m 'Add amazing feature')
4. Push to branch (git push origin feature/amazing)
5. Open Pull Request

---

📄 License

MIT License - Free for commercial and personal use!

See LICENSE for details.

---

🙏 Acknowledgments

• HuggingFace for transformers library
• Qwen Team for excellent models
• FAISS for fast vector search
• Sentence Transformers for embeddings
• Deep Translator for translation API
• pyttsx3 for text-to-speech

---

📧 Contact

• GitHub: @Giri530
• Email: [e-mail]
• Issues: Report bugs or request features

---

🌟 Star History

If you find OmniRAG useful, please ⭐ star the repo!

---

📚 Citation

@software{omnirag2025,
  title={OmniRAG: Multi-Language Voice RAG Framework},
  author={Girinath V},
  year={2025},
  version={2.0.0},
  url={https://github.com/Giri530/omnirag}
}

---

🎯 Roadmap

v2.0 (Current):

• ✅ Multi-language translation (27+ languages)
• ✅ Voice input and output
• ✅ UTF-8 encoding support

v2.1 (Planned):

• More file formats (DOCX, XLSX)
• Custom translation models
• Voice language selection
• GUI interface

v3.0 (Future):

• Real-time translation
• Multi-modal RAG (images)
• Cloud deployment
• API server

---

Made with ❤️ in India 🇮🇳

100% FREE Forever!

Happy RAG-ing! 🚀