📄🤖 ContextPDF: Multi-PDF Chat App

A Streamlit + LangChain + Gemini powered chatbot that lets you upload multiple PDFs and chat with them. The app extracts text, creates embeddings, stores them locally using FAISS, and answers your questions using RAG.

🚀 Features

• Upload multiple PDFs
• Chat with content using RAG
• Gemini 2.5 Flash for responses
• FAISS for fast vector search
• Automatic PDF hashing to avoid duplicate processing
• Local persistent vector store (multi_pdf_store/)
• Clean Streamlit UI with chat history & reset options

🛠️ Tech Stack

• Streamlit
• LangChain
• Google Gemini
• FAISS
• PyPDFLoader
• RecursiveCharacterTextSplitter

📦 Setup

1. Install dependencies

pip install -r requirements.txt

2. Run the app

streamlit run app.py

3. Configure API Key

Enter your Gemini API key in the sidebar or place it in .env as:

GOOGLE_API_KEY=your_key

📁 How It Works

1. PDFs → text extraction
2. Text → chunking
3. Chunk → embeddings (Gemini)
4. Stored in FAISS
5. RAG pipeline retrieves best chunks and generates an answer

🔧 Reset Options

• Clear Chat
• Reset All: removes chat, processed file hashes, and FAISS store