An end-to-end Retrieval-Augmented Generation (RAG) chatbot that lets you upload unstructured documents and ask questions in natural language, powered by Groq's Llama 3.3 70B for fast, context-aware response generation.
🔗 Live Demo: https://rag-chatbot-kr26.onrender.com/
Upload a document → Ask questions → Get intelligent, context-aware answers!
User uploads document ↓ LlamaIndex chunks & embeds the text ↓ Embeddings stored in ChromaDB (vector database) ↓ User asks a question ↓ Semantic search retrieves relevant chunks ↓ Groq (Llama 3.3 70B) generates a context-aware answer ↓ Answer streamed to a custom chat UI
- 📄 Upload and query unstructured documents
- 💬 Ask natural language questions with context-aware answers
- 🔍 Semantic search and retrieval via ChromaDB embeddings
- ⚡ Fast, cloud-hosted inference using Groq (Llama 3.3 70B)
- 🖥️ Custom streaming chat UI for real-time responses
- 🐳 Containerized with Docker and deployed on Render
| Component | Technology |
|---|---|
| LLM | Groq — Llama 3.3 70B (cloud-hosted) |
| RAG Framework | LlamaIndex |
| Vector Database | ChromaDB |
| Backend | FastAPI |
| UI | Custom streaming chat interface |
| Deployment | Docker, Render |
| Language | Python |
# 1. Clone the repo
git clone https://github.com/aralsaran36-cell/Qwen-AI-Assistant.git
cd Qwen-AI-Assistant
# 2. Install dependencies
pip install -r requirements.txt
# 3. Set your Groq API key
export GROQ_API_KEY=your_api_key_here
# 4. Run the backend
uvicorn app:app --reload- Chat with research papers
- Q&A on legal documents
- Study assistant for textbooks
- Internal knowledge base for companies
Saravanan S — Aspiring AI Engineer