MedTrack — Health Intelligence Platform

Clinical-grade AI that detects the Prescription Cascade before it harms you.

Built for patients, caregivers, and clinical decision support researchers.

Live Deployment: MedTrack

Overview

A Prescription Cascade happens when the side effect of one drug is mistaken for a new condition, prompting a second prescription — which creates its own side effects, leading to a third, and so on. Standard pharmacy software only checks two drugs at a time and completely misses these compounding chain reactions.

MedTrack is built to fix that.

It ingests your complete medication list, cross-references three drug-interaction databases (DDI, DrugBank, TWOSIDES), runs a multi-stage AI analysis, and calculates a clinical severity score — all in one place.

Key Features

AI Cascade Checker The core of MedTrack. Submit your medication list and the system runs a 3-tier LLM analysis powered by llama-3.3-70b-versatile via Groq. You get a brief side-effect summary, a contextual interaction paragraph, and a full human-readable clinical report — each escalating in depth.

Prescription OCR Scanner Upload a photo of any paper or digital prescription. The llama-4-scout-17b vision model reads the image, extracts every drug name and dosage, and identifies the likely prescribing specialist — with no manual entry required.

Voice Dictation Speak your medications aloud directly into the browser. The Web Speech API captures a live transcript and the backend AI extracts a structured medication list from your natural language.

Drug Library Browse a searchable A-Z encyclopedia of over 2,000 drugs sourced from multiple clinical datasets. Each entry includes known interactions, severity ratings, and cross-references.

Cloud History Sync Every analysis you run is saved instantly to browser localStorage and asynchronously backed up to Firestore. Your history is available across devices and is never lost.

PDF Report Export Generate a professionally formatted, branded PDF of any analysis — including the medication table, risk charts, and the full clinical report. Ready to share with your doctor.

Multi-language Analysis Switch the entire AI output between English, Hindi, and Marathi in real time. Designed to make clinical information accessible to a wider audience.

Read Aloud The Web Speech Synthesis API narrates the complete analysis report in the selected language — useful for accessibility and hands-free review.

Interactive Risk Dashboard A visual home base that shows animated safety and risk scores, a log of recent analyses, and live severity status panels — all powered by Recharts.

Images

Dashboard — Interactive Risk Overview

AI Cascade Checker — 3-Tier Analysis Report

Prescription OCR Scanner

Drug Library — A-Z Encyclopedia

Voice Dictation — Live Transcript

Architecture

flowchart TD
    subgraph Browser["🌐 Client Browser  (Next.js 16 · React 19 · TypeScript)"]
        direction LR
        UI["App UI\nDashboard · Checker · Library\nHistory · Profile"]
        FBSDK["Firebase SDK\nAuth + Firestore Client"]
    end

    subgraph Firebase["☁️ Firebase Cloud"]
        FAUTH["Firebase Authentication\nEmail / Password Sessions"]
        FSTORE["Firestore NoSQL DB\nUser History & Profiles"]
    end

    subgraph FlaskAPI["🐍 Flask Python API  (Port 5000)"]
        direction TB
        EP1["/api/check-interactions\nRun cascade analysis"]
        EP2["/api/scan-prescription\nOCR image extraction"]
        EP3["/api/extract-medications\nVoice transcript parsing"]
        EP4["/api/library/*\nA-Z drug encyclopedia"]
        SCORE["Severity Calculator\nDDI · DrugBank · TWOSIDES"]
        EP1 --> SCORE
    end

    subgraph Groq["🤖 Groq API"]
        LLM1["llama-3.3-70b-versatile\nCascade Analysis · Voice Extraction"]
        LLM2["llama-4-scout-17b\nPrescription Vision OCR"]
    end

    UI -->|"REST JSON"| FlaskAPI
    FBSDK <-->|"SDK"| Firebase
    FlaskAPI -->|"HTTPS"| Groq

Component Breakdown

Component	Functionality
Client Browser	Renders the Next.js 16 App Router UI. Handles client-side routing, state, and direct Firebase SDK calls for auth and history.
Firebase Authentication	Manages email/password login and signup. Persists user sessions in the browser with no server round-trip required.
Firestore DB	Stores user analysis history and profile data as NoSQL documents. Acts as the cloud backup in the dual-write system.
Flask API	Processes all backend logic — receives medication lists, queries the drug datasets, calculates algorithmic severity, and orchestrates Groq API calls.
Severity Calculator	Cross-references DDI, DrugBank, and TWOSIDES datasets to find pairwise interactions and compute a `Safe / Moderate / High / Critical` score.
`/api/check-interactions`	Receives the medication list + patient context, computes severity, then sends a structured prompt to `llama-3.3-70b` for the 3-tier analysis.
`/api/scan-prescription`	Accepts a base64-encoded prescription image and sends it to the `llama-4-scout-17b` vision model to extract drug names, doses, and specialist type.
`/api/extract-medications`	Receives a raw voice dictation transcript and uses `llama-3.3-70b` to parse it into a structured medication list.
*`/api/library/`**	Serves the full A-Z drug index and individual drug interaction records from the local DDI/DrugBank JSON database.
Groq API — llama-3.3-70b	Performs the core pharmacological reasoning: generates Short, Medium, and Detailed analysis reports and extracts medications from speech transcripts.
Groq API — llama-4-scout-17b	Handles multimodal vision inference — reads prescription images and returns structured JSON medication data.

Project Structure

Med_Track/
│
├── api/                           # Flask Backend
│   ├── index.py                   # Main API server (all endpoints)
│   ├── data_parser.py             # Drug dataset loader & query engine
│   ├── analyzer.py                # Algorithmic severity calculator
│   ├── drugs_database.json        # Local DDI drug database (JSON)
│   └── data/                      # Raw datasets (DrugBank, TWOSIDES)
│
├── src/                           # Next.js Frontend
│   ├── app/
│   │   ├── page.tsx               # Landing page
│   │   ├── layout.tsx             # Root layout (fonts, metadata)
│   │   ├── globals.css            # Global Tailwind styles
│   │   ├── login/                 # Login page
│   │   ├── signup/                # Signup page
│   │   └── dashboard/
│   │       ├── page.tsx           # Main dashboard
│   │       ├── layout.tsx         # Dashboard shell (Sidebar + TopBar)
│   │       ├── checker/           # AI Cascade Checker page
│   │       ├── history/           # Cloud history page
│   │       ├── library/           # Drug library page
│   │       └── profile/           # User profile page
│   │
│   ├── components/
│   │   └── dashboard/
│   │       ├── Sidebar.tsx        # Navigation sidebar
│   │       └── TopBar.tsx         # Header / topbar
│   │
│   └── lib/
│       ├── firebase.ts            # Firebase SDK initialization
│       ├── historyService.ts      # Dual-write history (localStorage + Firestore)
│       └── analysisService.ts     # API call wrappers for the Flask backend
│
├── requirements.txt               # Python dependencies (pinned versions)
├── package.json                   # Node dependencies
├── .env                           # Backend secrets (git-ignored)
├── .env.local                     # Frontend secrets (git-ignored)
├── next.config.ts                 # Next.js configuration
└── tsconfig.json                  # TypeScript configuration

Installation & Setup

Prerequisites

Requirement	Version	Notes
Node.js	≥ 18.x	nodejs.org
Python	≥ 3.10	python.org
Groq API Key	—	Free at console.groq.com
Firebase Project	—	console.firebase.google.com — enable Email/Password Auth and Firestore

Step 1 — Clone the Repository

git clone https://github.com/your-username/Med_Track.git
cd Med_Track

Step 2 — Backend Setup (Flask)

Run all commands from the project root (Med_Track/).

Create and activate a virtual environment:

# Windows (PowerShell)
python -m venv venv
.\venv\Scripts\activate

# macOS / Linux
python3 -m venv venv
source venv/bin/activate

Your terminal prompt will show (venv) when the environment is active.

Install Python dependencies:

pip install -r requirements.txt

Create the .env file in the project root:

# .env — backend secrets, never commit this file
GROQ_API_KEY=gsk_your_groq_api_key_here

Step 3 — Frontend Setup (Next.js)

Open a new terminal in the project root (keep the venv terminal open).

Install Node dependencies:

npm install

Create the .env.local file in the project root:

# .env.local — frontend secrets, never commit this file

# Flask backend URL
NEXT_PUBLIC_API_URL=http://localhost:5000

# Firebase — find these in Firebase Console → Project Settings → Your apps → Web App
NEXT_PUBLIC_FIREBASE_API_KEY=your_firebase_api_key
NEXT_PUBLIC_FIREBASE_AUTH_DOMAIN=your-project-id.firebaseapp.com
NEXT_PUBLIC_FIREBASE_PROJECT_ID=your-project-id
NEXT_PUBLIC_FIREBASE_STORAGE_BUCKET=your-project-id.firebasestorage.app
NEXT_PUBLIC_FIREBASE_MESSAGING_SENDER_ID=your_sender_id
NEXT_PUBLIC_FIREBASE_APP_ID=your_app_id

Running the Application

You need two terminals running at the same time:

Terminal	Command	URL
1 — Backend	`flask --app api/index run --port 5000`	http://localhost:5000
2 — Frontend	`npm run dev`	http://localhost:3000

Start the Flask backend first — the dashboard makes API calls on initial load.

Pre-flight checklist:

(venv) is active in Terminal 1
.env has a valid GROQ_API_KEY
.env.local has all 6 Firebase variables filled in
Firebase project has Email/Password authentication enabled
Firebase project has a Firestore database created (test or production mode)

Test Accounts

For ease of review and testing, you can log in using the following test credentials without creating a new account:

Account Type	Email	Password
Demo Patient	`patient@medtrack.demo`	`medtrack123`
Clinical User	`clinical@medtrack.demo`	`medtrack123`

Demo

Resource	Link
Video Demo	(Insert YouTube / Vimeo link here)
Voice Dictation Audio	(Insert audio file or link here)
Live Deployment	MedTrack

Tech Stack

Clinical Disclaimer

MedTrack is NOT a medical device.

This platform is an educational and clinical decision support tool only. It does not constitute medical advice, diagnosis, or treatment. The AI-generated analysis is intended to supplement — not replace — the professional judgment of a licensed physician or pharmacist.

Always consult a qualified healthcare professional before making any changes to your medication regimen.

Name		Name	Last commit message	Last commit date
api		api
public		public
scripts		scripts
src		src
.gitignore		.gitignore
.vercelignore		.vercelignore
AGENTS.md		AGENTS.md
CLAUDE.md		CLAUDE.md
README.md		README.md
add_languages.py		add_languages.py
apply_changes.py		apply_changes.py
eslint.config.mjs		eslint.config.mjs
next.config.ts		next.config.ts
package-lock.json		package-lock.json
package.json		package.json
postcss.config.mjs		postcss.config.mjs
requirements.txt		requirements.txt
test_api.py		test_api.py
test_api_paths.py		test_api_paths.py
tsconfig.json		tsconfig.json

Navigation Menu

Repository files navigation

MedTrack — Health Intelligence Platform

Table of Contents

Overview

Key Features

Images

Architecture

Component Breakdown

Project Structure

Installation & Setup

Prerequisites

Step 1 — Clone the Repository

Step 2 — Backend Setup (Flask)

Step 3 — Frontend Setup (Next.js)

Running the Application

Test Accounts

Demo

Tech Stack

Clinical Disclaimer

About

Releases

Packages

Contributors

Languages

Latest commit

History

Folders and files

Repository files navigation

MedTrack — Health Intelligence Platform

Table of Contents

Overview

Key Features

Images

Architecture

Component Breakdown

Project Structure

Installation & Setup

Prerequisites

Step 1 — Clone the Repository

Step 2 — Backend Setup (Flask)

Step 3 — Frontend Setup (Next.js)

Running the Application

Test Accounts

Demo

Tech Stack

Clinical Disclaimer

About

Resources

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages