Parallel RAG

A distributed RAG (Retrieval Augmented Generation) system that uses git-based parallelism for both development and runtime execution. Built on top of the Exo framework for robust topology, orchestration, and visualization capabilities.

Core Features

• Git-Based Parallelism: Uses git branches as a parallel execution mechanism
• Distributed RAG: Scalable retrieval augmented generation across nodes
• Vector Storage: Efficient similarity search and retrieval
• Z3 Solver Integration: Advanced reasoning capabilities
• Visualization: Real-time monitoring of parallel execution

Architecture

The system uses Exo's topology and orchestration components while adding:

1. Git-based parallel execution engine
2. Distributed vector storage
3. Z3 solver integration
4. Custom RAG implementation

Getting Started

pip install -e .

Usage

[Documentation coming soon]