Agent_ng

Vision

Agent_ng aims to implement modular, multi-agent orchestration for Retrieval-Augmented Generation (RAG) workflows. The project is designed for extensibility, reproducibility, and easy integration of new agents, tools, and data sources.

Features

• Modular RAG pipeline (embeddings, LLM, vector DB, loader)
• Multi-agent orchestration (planned)
• Streamlit UI for chat
• Chroma vector DB integration

Usage

1. Install dependencies:

   pip install -r requirements.txt
   

2. Run the app:

   python rag_app.py
   

3. Place data files in data/uploads/ for ingestion.

Roadmap

• Phase 0: Repo & environment setup
• Phase 1: Agent orchestration
• Phase 2: Tool integration
• Phase 3: Advanced workflows

Folder Structure

• core/: Main logic
• agents/: Agent extensions
• tools/: Tool extensions
• data/: Uploaded files and vector DB
• db/: Chroma DB storage

---

For architecture diagrams and roadmap, see the docs/ folder.