Metadata-Version: 2.1
Name: metarag
Version: 0.1.3
Summary: MetaRAG: A multi-LLM ensemble Retrieval-Augmented Generation framework with cosine similarity ranking.
Home-page: https://github.com/OpenRAG128/META-RAG
Author: Nisharg Nargund
Author-email: nisarg.nargund@gmail.com
Keywords: llm,retrieval,augmented,generation,ai,nlp
Classifier: Programming Language :: Python :: 3
Classifier: License :: OSI Approved :: MIT License
Classifier: Operating System :: OS Independent
Classifier: Development Status :: 4 - Beta
Classifier: Intended Audience :: Developers
Classifier: Intended Audience :: Science/Research
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Requires-Python: >=3.8
Description-Content-Type: text/markdown
Requires-Dist: langchain
Requires-Dist: langchain-groq
Requires-Dist: sentence-transformers
Requires-Dist: faiss-cpu
Requires-Dist: numpy
Requires-Dist: PyPDF2
Requires-Dist: scikit-learn

MetaRAG is a Python framework for multi-model Retrieval-Augmented Generation. It queries multiple LLMs in parallel, scores the responses based on cosine similarity with the context, and aggregates the top responses for a more accurate and comprehensive answer.

## Features
- 🔍 Multi-LLM querying using Groq's LLMs (LLaMA3, Gemma, etc.)
- 🤝 Cosine similarity scoring of responses
- 🧠 Top-k response aggregation
- 📄 Works with PDFs and plain text
- ⚡ Fast execution with thread pooling

## Installation
```bash
pip install metarag
```

## Example Usage
```python
from metarag import MetaRAG

rag = MetaRAG(["VectorDB_Paper.pdf"])
result = rag.query("Explain the abstract in simple terms")
print(result["aggregated_response"])
```

## Requirements
Python 3.8+

## License
MIT License - see LICENSE file for details.

## Author
**Nisharg Nargund**  
Founder @OpenRAG
