🧠 RAG Study Assistant

A full-stack Retrieval-Augmented Generation (RAG) study assistant that answers questions only from your uploaded content — no hallucinations, just grounded facts.

Built with:

• ⚛️ React (TypeScript)
• 🐹 Go (backend API proxy)
• 🐍 Python + FastAPI (RAG microservice)
• 🧠 LangChain + OpenAI
• 📦 Qdrant (vector DB for semantic search)

---

📸 Demo

Coming soon! (Optional: GIF or screenshot of app in action)

---

⚙️ Features

• Semantic document ingestion with LangChain
• Context-aware question answering using OpenAI
• Modular backend with Go -- FastAPI HTTP bridge
• Answers grounded in your own files (notes, PDFs, etc.)
• Returns “I don’t know” when no relevant context is found to avoid hallucination
• Fast and lightweight — runs locally with Docker + Vectors

---

🧰 Tech Stack

Layer	Technology
Frontend	React (TypeScript, Next.js)
Backend	Go (API proxy → Python microservice)
Microservice	FastAPI, LangChain, OpenAI
Vector Store	Qdrant (Dockerized)

---

🚀 Getting Started

Prerequisites:

• Docker + Docker Compose
• Node.js + npm
• Python 3.10+
• Go 1.21+

Open separate terminal windows (Command Prompt recommended) for each of the following commands:

1. Start Qdrant (Vector DB)

cd qdrant docker-compose up -d

2. Start FastAPI (Python RAG Microservice)

cd .. uvicorn rag_pipeline.api:app --reload --port 8000

3. Start Go Backend

cd backend go run main.go

4. Start React Frontend

cd ../frontend npm install # Only needed once npm run dev

5. Ingest Your Documents (Required Before Querying)

cd ../rag_pipeline python ingest.py path/to/your/file.txt