Project Name: RAG Powered News Chatbot

Description: This project is a backend service that implements a Retrieval-Augmented Generation (RAG) chatbot. The chatbot answers user queries by retrieving relevant information from a news dataset and generating accurate responses using a Large Language Model (Google Gemini).

The system first searches a vector database (Qdrant) to find relevant news content and then passes that context to the LLM to produce reliable, up-to-date answers.

Key Features:

• Retrieval-Augmented Generation (RAG)
• News-based question answering
• Google Gemini LLM integration
• Vector search using Qdrant
• Session-based chat support
• REST API for frontend integration

Tech Stack:

• Node.js
• Express.js
• Google Gemini API
• Qdrant Vector Database
• Redis (optional for sessions)

Use Case: Designed for answering questions over news articles with higher accuracy and reduced hallucinations compared to traditional LLM-only chatbots.

backend/ ├── src/ │ ├── routes/ │ │ └── chat.routes.js │ ├── controllers/ │ │ └── chat.controller.js │ ├── services/ │ │ ├── gemini.service.js │ │ ├── embedding.service.js │ │ ├── rag.service.js │ │ └── qdrant.service.js │ ├── utils/ │ │ └── chunkText.js │ ├── data/ │ │ └── news.json │ └── app.js ├── .env ├── package.json └── README.md