An AI-powered computer vision platform built with Next.js, TensorFlow.js, MobileNet, and COCO-SSD for Image Classification and Real-Time Object Detection—all running directly in your browser.
🚀 Live Application: https://vision-ai-henna.vercel.app/
VisionAI Studio is a modern AI-powered web application that demonstrates the capabilities of browser-based computer vision using TensorFlow.js.
The application provides two major AI modules:
- 🖼️ Image Classification
- 🎥 Real-Time Object Detection
Unlike cloud-based AI applications, VisionAI Studio performs inference directly inside the browser, ensuring better privacy and lower latency.
- Upload an image
- AI predicts image content using MobileNet
- Displays Top-3 predictions
- Confidence score visualization
- Beautiful responsive UI
- Webcam-based object detection
- Detects 80+ everyday objects
- Bounding boxes around detected objects
- Live confidence scores
- FPS counter
- AI status indicators
- Dark futuristic interface
- Glassmorphism cards
- Fully responsive
- Mobile friendly
- Smooth animations
- Gradient design
All AI inference happens locally in your browser.
✅ No images are uploaded.
✅ Webcam feed never leaves your device.
- Next.js
- React
- Tailwind CSS
- TensorFlow.js
- MobileNet
- COCO SSD
- Lucide React
- React Icons
src
│
├── app
│ ├── page.jsx # Home
│ ├── about
│ ├── image-classifier
│ ├── live-detection
│ ├── layout.js
│ └── globals.css
│
├── components
│ ├── Navbar.jsx
│ └── ...
│
└── public
Clone the repository
git clone https://github.com/SoyebAhmed5/VisionAI.gitGo to the project
cd visionaiInstall dependencies
npm installRun locally
npm run devOpen
http://localhost:3000
Used for
- Image Classification
- Top Predictions
- Confidence Scores
Used for
- Real-Time Detection
- Multiple Object Detection
- Bounding Boxes
- Confidence Scores
- Face Detection
- Face Recognition
- Emotion Detection
- OCR (Text Recognition)
- Pose Estimation
- QR Code Scanner
- Image Segmentation
- YOLO Integration
- Custom Model Upload
- Object Tracking
- Browser-Based AI
- Real-Time Inference
- No Backend Required
- Lightweight
- Responsive UI
✔ Runs entirely in the browser
✔ Fast AI predictions
✔ Privacy-friendly
✔ Beautiful modern interface
✔ Responsive design
✔ Built using modern web technologies
Soyeb Ahmed
Full Stack Developer • AI Enthusiast • Computer Science Student
- LinkedIn: https://www.linkedin.com/in/soyebahmed2005
- GitHub: https://github.com/SoyebAhmed5
Live Website
https://vision-ai-henna.vercel.app/
This project is licensed under the MIT License.
If you found this project useful, please consider giving it a ⭐ on GitHub!