MedRemind: Personal Medicine & Health Log

Full Stack Hackathon Project Built with React, Node.js, MongoDB, OpenAI, and Pinecone (2-Day Solo Build).

Problem Statement

People often forget to take their medications on time and rarely keep track of their daily health habits. There is no simple, unified tool that allows individuals to manage their medicines, log daily health metrics, and get instant guidance for common health issues all in one place.

MedRemind is a personal medicine and health log web application that lets users add and manage their medications, log daily health notes (mood, weight, blood pressure), view their health history with filtering and pagination, and interact with an AI-powered chatbot that suggests common medicines for everyday symptoms like fever, gastric issues, headache, and more, powered by RAG (Retrieval-Augmented Generation) and a vector database.

---

Tech Stack

• Frontend: ReactJS + Tailwind CSS + React Router
• Backend: Node.js + Express.js
• Database: MongoDB (Mongoose)
• Vector DB: Pinecone (free tier)
• Embeddings: OpenAI text-embedding-ada-002
• LLM: OpenAI GPT-3.5-turbo
• RAG Layer: LangChain.js

---

Mandatory Features Coverage

Feature	Implementation in MedRemind
Routing & Navigation	Home, Login, Signup, Dashboard, Medicines, Health Log, Profile
React Hooks	useState, useEffect, useRef, useContext — all demonstrated
State Management	Context API for auth state + theme state
Authentication	Signup / Login with LocalStorage + Protected Routes
Theme Support	Dark / Light mode toggle, preference persisted
Search, Filter, Sort	Search by medicine name; filter by date/category; sort by time
Pagination	Medicine list + Health log history (backend limit/skip)
CRUD Operations	Create / Read / Update / Delete medicines & health logs
API Integration	REST APIs via Express; loading states; error handling
Form Handling	Controlled components, input validation, error messages
Responsive UI	Tailwind CSS—mobile, tablet, desktop layouts
Error Handling	Try-catch on all API calls; backend error responses

---

MongoDB Schema

• Users: {_id, name, email, password, createdAt}
• Medicines: {_id, userId, name, dosage, frequency, category, startDate, endDate, notes, createdAt}
• HealthLogs: {_id, userId, date, mood, weight, bloodPressure, notes, createdAt}

---

API Routes

Method	Endpoint	Purpose
POST	/api/auth/signup	Register user
POST	/api/auth/login	Login user
GET	/api/medicines	Get all medicines (search, filter, paginate)
POST	/api/medicines	Add medicine
PUT	/api/medicines/:id	Update medicine
DELETE	/api/medicines/:id	Delete medicine
GET	/api/logs	Get health logs (filter by date, paginate)
POST	/api/logs	Add health log
PUT	/api/logs/:id	Update health log
DELETE	/api/logs/:id	Delete health log
POST	/api/chat	RAG Chatbot: symptom to medicine suggestion

---

AI Chatbot with RAG + Vector Database

MedRemind integrates an AI-powered medical assistant chatbot that uses Retrieval-Augmented Generation (RAG) to suggest appropriate medicines for common everyday symptoms such as fever, gastric issues, headache, cold, and acidity. The chatbot does not rely on model memory alone; it retrieves accurate, structured knowledge from a curated medical knowledge base stored in a vector database.

RAG Architecture Flow

1. User Input: User types a symptom e.g., 'I have fever and body ache'.
2. Embedding: Query is embedded using OpenAI text-embedding-ada-002.
3. Vector Search: Pinecone finds top-3 most similar knowledge chunks.
4. Prompt Building: Retrieved context + user query form the LLM prompt.
5. LLM Response: GPT-3.5-turbo returns medicine name, dosage & precautions.
6. Chat UI: Response shown in real-time chat bubble interface.

RAG Components

Component	Technology	Purpose
Vector Database	Pinecone	Free tier, cloud-hosted, fast similarity search
Embeddings	OpenAI text-embedding-ada-002	Converts text to 1536-dim vectors
LLM	GPT-3.5-turbo	Generates medicine suggestions from context
RAG Framework	LangChain.js	Orchestrates retrieval + prompt + LLM call
Knowledge Base	.txt files per condition	fever, gastric, headache, cold, acidity, etc.

Knowledge Base Example (gastric.txt): Condition: Gastric / Gas & Bloating Symptoms bloating, stomach pain, burping, flatulence, indigestion Common Medicines: Pantoprazole 40mg (before meals, once daily) Domperidone 10mg (after meals) Digene syrup (antacid, after meals) Precautions: Avoid spicy food, eat small meals. See a doctor if pain is severe or lasts more than 3 days.

---