🤖 Kore-Ai: Intelligent RAG Document Comparison

Kore-Ai is a sophisticated Retrieval-Augmented Generation (RAG) chatbot built to solve the challenge of manual document review. It allows users to upload multiple PDF documents, ingests them into a vector database, and uses local LLMs to perform detailed comparison, summarization, and gap analysis in real-time.

🚀 Key Features

• Dual-Document Ingestion: Upload and analyze two distinct PDF documents simultaneously (e.g., Document A vs. Document B).
• Semantic Search: Uses MongoDB Atlas Vector Search to retrieve context based on meaning, not just keywords.
• Local AI Power: Privacy-focused architecture using Ollama and Llama 3.2 running locally.
• Persistent Memory: All chat history and user interactions are securely stored in MongoDB, allowing for long-term context retention.
• Real-Time Streaming: Fast, typewriter-style responses for a smooth user experience.
• Modern UI: Fully responsive, dark-themed interface built with React and Tailwind CSS.

🛠️ Tech Stack

Component	Technology	Description
Frontend	React.js, Tailwind CSS	Responsive UI for document upload and chat.
Backend	Python (Flask/FastAPI)	API server handling ingestion and LLM requests.
Database	MongoDB Atlas	Stores Vector Embeddings (document_chunks) & Chat History.
AI Model	Llama 3.2 (via Ollama)	Lightweight, high-speed local LLM for reasoning.
Embeddings	Nomic-Embed-Text	Converts PDF text into vector representations.

⚙️ Prerequisites

Before running the project, ensure you have the following installed:

1. Python 3.10+
2. Node.js & npm
3. Ollama (Running locally)
4. MongoDB Atlas Account (With a dedicated Cluster)

📦 Installation & Setup

1. Clone the Repository

git clone [https://github.com/Adesh2204/Kore-Ai.git](https://github.com/Adesh2204/Kore-Ai.git)
cd Kore-Ai

2. Backend Setup

Navigate to the backend directory:

# Create a virtual environment
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

🔐 Security Configuration (Important): Create a file named .env in the root directory. This file is not tracked by git to protect your secrets. Add your specific credentials:

MONGO_URI=your_mongodb_connection_string_here
DB_NAME=rag_chatbot
COLLECTION_NAME=document_chunks

⚠️ Note: Never commit your .env file to GitHub. Ensure it is listed in your .gitignore file.

3. Frontend Setup

Navigate to the frontend folder:

cd frontend
npm install

4. AI Model Setup

Ensure Ollama is running and pull the required models:

ollama pull llama3.2
ollama pull nomic-embed-text

🏃‍♂️ How to Run

Step 1: Start the Backend Server

# In your backend terminal
python api_server.py

Step 2: Start the Frontend Client

# In your frontend terminal
npm run dev

Step 3: Access the App Open your browser and navigate to http://localhost:5173 (or the port shown in your terminal).

📂 Project Structure

Kore-Ai/
├── frontend/          # React Source Code
├── uploads/           # Temporary storage for uploaded PDFs
├── api_server.py      # Main Backend API entry point
├── ingest.py          # Script for chunking & vectorizing PDFs
├── requirements.txt   # Python dependencies
└── .gitignore         # Ensures secrets (.env) are not uploaded

🤝 Contributing

Contributions are welcome! Please fork the repository and submit a pull request for any enhancements.

📄 License

This project is open-source and available under the MIT License.