📄 Chat with Your PDF - RAG Application Transform any PDF into an interactive conversation! Upload your documents and ask questions in natural language powered by cutting-edge AI technology. Show Image Show Image Show Image Show Image 🌟 Live Demo Try it here! (Add your Streamlit Cloud link after deployment) ✨ Features
📤 Upload Any PDF - Research papers, books, reports, manuals 💬 Natural Conversations - Ask questions like you're talking to a human ⚡ Lightning Fast - Powered by Groq's ultra-fast LLM inference 🎯 Accurate Answers - RAG ensures responses are grounded in your document 🎨 Beautiful Interface - Clean, intuitive design built with Streamlit 🔒 Privacy First - Process documents locally, no data stored
Demo Link : https://pdf-chat-rag-fx5nczbrwczzpou6qyczmj.streamlit.app/ (Note: You'll need a free Groq API key to chat — sign up in seconds at https://console.groq.com 🔥)
🛠️ Tech Stack TechnologyPurposeWhy We Chose ItLangChainRAG FrameworkIndustry-standard for building LLM apps with document retrievalGroqLLM Inference10x faster than traditional inference, free tier availableFAISSVector SearchFacebook's efficient similarity search libraryStreamlitUI FrameworkRapid prototyping with beautiful Python-based interfacesHuggingFaceEmbeddingsOpen-source sentence transformers for text vectorization 🚀 Quick Start Prerequisites
Python 3.8 or higher Groq API key (Get one free here)
Installation
Clone the repository
bashgit clone https://github.com/aliabdm/pdf-chat-rag.git cd pdf-chat-rag
Create virtual environment
bashpython -m venv venv
venv\Scripts\activate
source venv/bin/activate
Install dependencies
bashpip install -r requirements.txt
Run the application
bashstreamlit run app.py
Open your browser Navigate to http://localhost:8501
📖 How It Works The RAG Pipeline PDF Upload → Text Extraction → Chunking → Embedding → Vector Store ↓ User Question → Embedding → Similarity Search → Context Retrieval ↓ Context + Question → LLM → Answer Step-by-Step Process
Document Processing
Extract text from PDF using PyPDF2 Split into manageable chunks (1000 chars with 200 overlap) Convert chunks to vector embeddings using HuggingFace
Vector Storage
Store embeddings in FAISS for fast similarity search Enable quick retrieval of relevant document sections
Query Processing
Convert user question to embedding Find top-3 most similar document chunks Pass context to LLM with original question
Answer Generation
Groq's LLM generates contextual response Answers grounded in document content Fast inference (typically under 2 seconds)
💡 Example Use Cases Academic Research "What are the main findings of this research paper?" Legal Documents "What does section 5 say about liability?" Technical Manuals "How do I troubleshoot error code E404?" Business Reports "Summarize Q3 financial performance" 🔧 Configuration Customize Chunk Size pythontext_splitter = RecursiveCharacterTextSplitter( chunk_size=1000, # Adjust based on your documents chunk_overlap=200 # Balance between context and uniqueness ) Change LLM Model pythonllm = ChatGroq( model_name="llama-3.3-70b-versatile", # Try other Groq models temperature=0 # 0 for factual, higher for creative ) Adjust Retrieval pythondocs = vector_store.similarity_search( question, k=3 # Number of chunks to retrieve ) 📊 Performance
Processing Time: ~10-30 seconds for typical PDFs Query Response: ~1-3 seconds per question Memory Usage: Scales with document size Accuracy: High for factual questions, context-dependent
🎯 Roadmap
Support for multiple file formats (DOCX, TXT, etc.) Conversation history persistence Multi-document chat Export chat transcripts Advanced filtering options API endpoint for integration
🤝 Contributing Contributions are welcome! Feel free to:
Fork the repository Create a feature branch (git checkout -b feature/AmazingFeature) Commit your changes (git commit -m 'Add some AmazingFeature') Push to the branch (git push origin feature/AmazingFeature) Open a Pull Request
📝 License This project is licensed under the MIT License - see the LICENSE file for details. 🙏 Acknowledgments
LangChain for the powerful RAG framework Groq for ultra-fast LLM inference Streamlit for the amazing UI framework Facebook AI for FAISS vector search HuggingFace for open-source embeddings
📧 Contact Mohammad Ali Abdul Wahed
⭐ If you find this project helpful, please give it a star! ⭐ Built with ❤️ using Python, LangChain, and Groq