A RAG-based AI study assistant designed for university students. It allows students to ask questions from their study materials and receive AI-generated answers based on the relevant content retrieved from uploaded documents.
- 📚 PDF Study Material Support
- 🔎 Semantic Search using RAG
- 🤖 AI-generated answers
- 📄 Source references with page numbers
- 🧠 Hugging Face local AI models
- ⚡ FastAPI backend
- 💻 Next.js + TypeScript frontend
- 🗄️ Qdrant vector database
- 🐳 Docker & Docker Compose
- 🔐 No OpenAI API required
| Technology | Purpose |
|---|---|
| Next.js | Frontend |
| TypeScript | Frontend development |
| Tailwind CSS | UI styling |
| Python | Backend & RAG |
| FastAPI | REST API |
| Sentence Transformers | Text embeddings |
| Qdrant | Vector database |
| Hugging Face | Local LLM |
| PyTorch | Model inference |
| Docker | Containerization |
University PDFs
↓
PDF Loader
↓
Cleaner
↓
Chunker
↓
Embeddings
↓
Qdrant Vector DB
↓
User Question
↓
Question Embedding
↓
Semantic Search
↓
Relevant Documents
↓
Local LLM
↓
AI Answer
↓
Answer + Sources
University-AI-Study-Assistant/
│
├── backend/
│ ├── app/
│ │ ├── rag/
│ │ │ ├── loader.py
│ │ │ ├── cleaner.py
│ │ │ ├── chunker.py
│ │ │ ├── embeddings.py
│ │ │ ├── vector_store.py
│ │ │ ├── generator.py
│ │ │ └── pipeline.py
│ │ ├── main.py
│ │ └── __init__.py
│ │
│ ├── requirements.txt
│ └── Dockerfile
│
├── frontend/
│ ├── app/
│ ├── public/
│ ├── package.json
│ └── Dockerfile
│
├── data/
│ └── documents/
│
├── docker-compose.yml
├── .gitignore
└── README.md
Clone the repository:
git clone https://github.com/Shintoo-Engineer/University-AI-Study-Assistant.gitGo into the project:
cd University-AI-Study-AssistantStart the application:
docker compose up -d --buildThe application will run on:
Frontend: http://localhost:3001
Backend: http://localhost:8000
GET /healthPOST /askExample:
{
"question": "What is Docker?"
}The API returns the generated answer along with the retrieved document sources and page numbers.
The goal of this project is to provide university students with an AI-powered learning assistant that can answer questions from their actual study materials rather than relying only on general-purpose AI knowledge.
- ✅ PDF ingestion
- ✅ Text cleaning
- ✅ Text chunking
- ✅ Embeddings
- ✅ Qdrant vector search
- ✅ RAG retrieval
- ✅ Local LLM generation
- ✅ FastAPI API
- ✅ Next.js frontend
- ✅ Docker Compose
- 🚧 Public cloud deployment
University AI Study Assistant
Built as a university-focused RAG application for intelligent study-material search and question answering.