A Streamlit-based application allowing users to upload one or multiple PDFs, then interact with them through a chat interface. It intelligently extracts and chunks PDF text, generates embeddings via Google’s embedding model, and creates a FAISS vector store for similarity search. User queries fetch the top relevant text chunks, which are then used by a generative AI to produce context-aware answers.
Key Features
Upload and interact with single or multiple PDF documents via Streamlit
Text extraction and smart chunking of PDF content
Semantic embedding using Google Generative AI model ("embedding‑001")
Efficient retrieval via FAISS vector store
Conversational flow with context-awareness and chat history
Retrieval + generation pipeline (RAG) for accurate responses
Easy setup via requirements.txt and .env for Google API key
Android DevHub
.png)
.png)
.png)
