Vriksha Vaidya is a professional-grade, browser-native application designed to empower farmers and botanists with instantaneous plant disease diagnostics. Powered by an ultra-optimized MobileNet neural network running completely offline via TensorFlow.js, it can accurately detect up to 38 distinct plant diseases securely on your device.
Featuring a cutting-edge 3D Glassmorphic UI built with WebGL and native CSS, Vriksha Vaidya delivers both blazing-fast AI inference and a visually stunning user experience—all without relying on a backend server.
- 🧠 Zero-Latency Edge AI: Real-time inference using TensorFlow.js (WASM / WebGL hardware acceleration).
- 🔒 100% Privacy-First: No images or data are ever uploaded to a server. Inference happens exclusively in your browser.
- 🌿 Broad Botanical Coverage: Detects 38 different classes of plant conditions (healthy and diseased) across multiple crop species.
- 📱 Progressive & Responsive: Fully functional on mobile devices, tablets, and desktops with deep touch gesture support.
- 🌌 Interactive 3D UI: Engineered with Three.js environmental physics, deep CSS parallax rendering, and Vanilla-Tilt dynamics.
- 📚 Disease Encyclopedia: Comprehensive offline database providing symptoms, causes, and actionable treatment recommendations.
- 🗂️ Local Storage History: Automatically saves and tracks your past scans in the browser via IndexedDB and LocalStorage for future reference.
Vriksha Vaidya is built to demonstrate the immense power of the modern open web, utilizing a pure vanilla stack strictly enforcing Separation of Concerns. No heavy frontend frameworks. No backend APIs.
- Core Engine: Vanilla JavaScript (ES6 Modules)
- AI & Machine Learning: TensorFlow.js (
@tensorflow/tfjs) - Styling & Layout: Modular Vanilla CSS3 (Custom Variables, Flexbox, Grid, Advanced Keyframes)
- 3D Visuals & Physics: Three.js (r128), Vanilla-Tilt.js
- Storage: HTML5 LocalStorage & IndexedDB API
Because Vriksha Vaidya is entirely client-side, running it locally requires zero build steps or package managers.
A modern web browser (Chrome, Edge, Firefox, Safari) and a simple local server to bypass standard CORS module restrictions.
-
Clone the repository:
git clone https://github.com/abhranilsingharoy-cloud/Vriksha-Vaidya.git cd Vriksha-Vaidya -
Serve locally: You can use any local server. If you have Python installed:
python -m http.server 8080
Or using Node.js (
npx):npx serve . -
Open the App: Navigate to
http://localhost:8080in your browser.
Vriksha-Vaidya/
├── index.html # Main application entry point
├── assets/ # Icons, SVGs, and brand images
├── css/ # Modular styling system
│ ├── animations.css # 3D transforms & CSS Keyframes
│ ├── components.css # Buttons, badges, and shared UI elements
│ ├── layout.css # Grid layouts and Glassmorphic panels
│ └── ... # Section-specific stylesheets
└── js/ # Core Application Logic
├── config.js # Global configuration and AI Model mapping
├── main.js # Application bootstrapper
├── model-loader.js # TensorFlow.js initialization and IDB caching
├── inference.js # Vision tensor manipulation and forward pass logic
├── three-scene.js # Three.js WebGL terrain rendering
└── ... # Isolated modular logic (history, scanner, etc.)
This project is inherently designed as a static site and is ready to be deployed to any static hosting provider like Vercel, Netlify, or GitHub Pages.
For Vercel:
- Import your GitHub repository to Vercel.
- Select "Other" as the Framework Preset.
- Click Deploy. Vercel will instantly host your app globally!
Distributed under the MIT License. See LICENSE for more information.
Abhranil Singha Roy
Passionate about bridging the gap between cutting-edge Artificial Intelligence and practical, accessible tools for the real world.