Medical AI Chatbot

Medical AI Chatbot is a final year project focused on building an intelligent, retrieval-augmented assistant for medical question answering. The system combines a modern web interface, a FastAPI backend, vector search, medical document preprocessing, and RAG evaluation workflows to support contextual responses from curated medical knowledge sources.

This project was built by Kiran Seenivasan, Isiri HV, Yashaswini P, and Anubhav Kushagra from the 2026 batch of Computer Science and Business Systems (CSBS).

Overview

The repository contains the main chatbot application and supporting research workflows used for medical document processing, embedding generation, vector storage, retrieval evaluation, and model experimentation.

The goal of the project is to make medical information retrieval more accessible through a conversational interface. The chatbot is designed to retrieve relevant context from indexed medical documents and use that context to generate more grounded responses.

Note: This project is intended for academic and research purposes. It is not a replacement for professional medical advice, diagnosis, or treatment.

Key Features

• Medical AI chatbot interface built with React and TypeScript.
• FastAPI backend for chatbot and retrieval workflows.
• Retrieval-Augmented Generation (RAG) pipeline for context-aware answers.
• Qdrant-based vector search experiments and storage workflows.
• Medical document preprocessing and chunking notebooks.
• Embedding and retrieval evaluation notebooks.
• Modular backend logic for integration with vector databases and model pipelines.

Repository Structure

.
├── frontend/
│   ├── FYP/
│   │   ├── fastapi-backend/      # Main backend API
│   │   └── react-frontend/       # Main chatbot frontend
│   └── Medocembed/               # Embedding and Qdrant helper logic
├── Qdrant/                       # Vector database, preprocessing, and RAG experiments
├── mdbIE/                        # Additional RAG package experiments
└── .gitignore                    # Project-wide ignore rules

Tech Stack

• Frontend: React, TypeScript, Vite, Tailwind CSS
• Backend: FastAPI, Python
• Vector Database: Qdrant
• AI/RAG: Embeddings, semantic retrieval, document chunking, RAG evaluation
• Tooling: Jupyter notebooks, npm, pip

Getting Started

Frontend

cd frontend/FYP/react-frontend
npm install
npm run dev

Backend

cd frontend/FYP/fastapi-backend
python -m venv venv
source venv/bin/activate
pip install -r requirements.txt
python main.py

Qdrant

The repository includes Qdrant experiment files and a Docker Compose configuration:

cd Qdrant
docker compose up -d

Generated Qdrant storage, model weights, caches, logs, virtual environments, and local environment files are ignored by Git.

Contributors

• Kiran Seenivasan
• Isiri HV
• Yashaswini P
• Anubhav Kushagra

Academic Details

• Batch: 2026
• Department: Computer Science and Business Systems (CSBS)
• Project Type: Final Year Project

License

This project is licensed under the MIT License.

Copyright (c) 2026 Kiran Seenivasan, Isiri HV, Yashaswini P, and Anubhav Kushagra

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files, to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, subject to the conditions of the MIT License.