ResearchAI

An AI-powered research assistant built with Django and Llama 3.1.

What it does

• Enter any topic and get a structured research summary instantly
• AI remembers your past research, connects and provides related topics (RAG)
• Personal research history saved per user
• Secure authentication system

Tech Stack

Backend

• Django + Django REST Framework
• SQLite database
• ChromaDB vector database

AI

• Llama 3.1 via Groq API
• RAG (Retrieval Augmented Generation)
• Sentence Transformers for embeddings

Frontend

• HTML, CSS, Bootstrap 5
• Vanilla JavaScript (fetch API)

Architecture

User submits topic
→ Django checks ChromaDB for similar past research (RAG)
→ Sends topic + context to Llama 3.1 via Groq
→ AI generates structured research summary
→ Result saved to SQLite + ChromaDB
→ Displayed to user instantly

Setup

1. Clone the repo
2. Create virtual environment python -m venv venv venv\Scripts\activate
3. Install dependencies pip install -r requirements.txt
4. Create .env file SECRET_KEY=your_django_secret_key DEBUG=True GROQ_API_KEY=your_groq_api_key
5. Run migrations python manage.py migrate
6. Start server python manage.py runserver

Key Concepts Implemented

• Custom Django User model with email authentication
• REST API endpoints with DRF
• RAG pipeline using ChromaDB + Sentence Transformers
• Agentic AI with context-aware responses
• Secure API key management with python-dotenv

Built With

• bootstrap
• chromadb
• css
• django
• groq
• html
• python
• sqlite