📚 Streamlit RAG Chatbot App with Groq
This project is a Retrieval-Augmented Generation (RAG) pipeline built with Streamlit, LangChain, FAISS, and Groq LLMs. It allows you to upload PDF documents, embed and store them, and query the content using a powerful language model backend.
🚀 Features
Upload and parse PDF documents with PyMuPDF (fitz)
Split text using RecursiveCharacterTextSplitter
Generate vector embeddings with HuggingFaceEmbeddings
Store and query embeddings in a FAISS vector database
Ask natural language questions and get context-aware answers via Groq LLMs
Simple, interactive Streamlit web UI
🛠️ Tech Stack
Streamlit – UI framework
LangChain – RAG pipeline utilities
FAISS – Vector search
HuggingFace Embeddings – Embeddings
Groq – LLM inference
PyMuPDF – PDF parsing
📂 Project Structure
.
├── app.py # Main Streamlit app
├── requirements.txt # Dependencies
├── .env # API keys (Groq, HuggingFace, etc.)
└── vectorstore/ # FAISS index (auto-created after upload)
- Add environment variables
Create a .env file:
GROQ_API_KEY=your_groq_api_key HF_MODEL=sentence-transformers/all-MiniLM-L6-v2
- Run the app streamlit run app.py
🎯 Usage
Open the Streamlit app in your browser
Upload one or more PDF files
Ask questions about the documents
Get AI-powered answers with citations from your PDFs
🔮 Future Improvements
Add support for multiple embedding models
Implement persistent vector storage (e.g., Pinecone, Weaviate)
Add document summarization and chat history