Med AI

Project Overview

A unified healthcare AI solution providing instant drug analysis, pill identification, and symptom triage via a premium Web Dashboard and a lightweight Telegram Bot.

The Challenge

Handling complex data synchronization across multiple devices without internet.

The Solution

Custom conflict-resolution logic with local IndexedDB storage.

Technical Deep Dive

1. Executive Summary

MediBot AI is a unified healthcare solution designed to simplify access to complex medical information. By combining the power of Google Gemini AI and OpenFDA, it provides users with instant drug analysis, pill identification, and symptom triage across two high-accessibility platforms: a premium Glassmorphism web dashboard and a lightweight Telegram conversational bot.

2. Problem Statement

The modern patient is often overwhelmed by medical data that is either too technical or geographically limited:

• The "Jargon" Barrier: Official FDA labels are dense and written for medical professionals.

• The "International Gap": Global users often have local brand names (e.g., Dolo 500) that are not recognized by US-centric databases.

• Safety Fragmentation: Tools for pill identification and interaction checking are usually separate.

• Accessibility Needs: Users in high-stress situations need simple, text-based advice without loading heavy web pages.

3. Technology Stack

The system uses an Industry-Grade stack built for security, speed, and cross-platform intelligence.

• Backend: Node.js with Express.js (High-performance event-driven architecture).

• Database: SQLite (Lightweight, local storage for Medicine Cabinet persistence).

• Security: Helmet.js (HTTP header security) and express-rate-limit.

• AI Engine: Google Gemini 2.0 Flash (Multimodal AI).

• Medical Source: OpenFDA API (Real-time connection to official US government data).

• Vision AI: Google Generative AI Vision (for identifying pill images).

4. Key Features

Section A: The Web Platform

• AI Bridge Search: Translates international brands into generic names to unlock FDA data.

• Interactive AI Analysis: Summarizes complex drug documents into clear Common Use and Warning sections.

• Visual Pill ID: Upload photos to identify unknown tablets using AI Vision.

• Floating Chat Window: A draggable and resizable AI chatbot.

Section B: Telegram Bot (@MedicalDrugInfo_Bot)

• Instant Text Search: Quick identification via simple text commands.

• Photo Processing: Send a photo of a medicine box for instant AI identification.

• Mobile-First Design: Optimized for low internet connectivity.

Section C: Safety Tools

• Interaction Safety Engine: Cross-references two drugs to alert users of dangerous combinations.

• Symptom Triage Assistant: Acts as a virtual nurse identifying potential issues.

• SOS Mode: One-click access to emergency contact numbers.