Nykaa-Annual-Report-RAG-Chatbot

Live Demo :- https://nykaa-annual-report-rag-chatbot-6bsuzz2ngf6tbkmhcdsrzm.streamlit.app/

A Retrieval-Augmented Generation (RAG) system built using LangChain, ChromaDB, HuggingFace Embeddings, and a local free LLM (FLAN-T5).

Features

Extracts insights from Nykaa’s 2024–25 Annual Report

Embeds PDF using free all-MiniLM-L6-v2 embeddings

Uses local Flan-T5 model for answering

Structured and accurate outputs

100% offline (no OpenAI cost)

Tech Stack

Python

LangChain

ChromaDB

HuggingFace Transformers

SentenceTransformers

Jupyter Notebook

Future Improvements

Replace FLAN-T5 with Mistral or Llama models

Add Streamlit UI

Build a fully deployed web chatbot

Built With

• jupyter-notebook
• python