This project implements a Retrieval-Augmented Generation (RAG) system that allows users to upload documents in various formats such as PDF, Word (DOCX), TXT, and more. The contents of these documents are processed and stored as vector embeddings in ChromaDB, a vector database. Using these embeddings, the system enables efficient retrieval of relevant information to augment the generation of responses by a chatbot interface, built with Gradio.
- ✅ Supports multiple document formats: PDF, DOCX, TXT, etc.
- ✅ Text extraction and preprocessing pipeline.
- ✅ Embedding generation using state-of-the-art models.
- ✅ Vector storage and retrieval with ChromaDB.
- ✅ Conversational chatbot interface powered by Gradio.
- ✅ Retrieval-Augmented Generation for more accurate, context-aware responses.
- Python
- LangChain or LlamaIndex (for RAG orchestration) — optional but recommended
- ChromaDB (Vector Database)
- Sentence Transformers / OpenAI embeddings (for vector representation)
- Gradio (for chatbot interface)
- PyMuPDF, python-docx (for document parsing)