ImageKit | AI Image Helper Kit

ImageKit is a high-performance, privacy-focused image processing suite built with Node.js & Express. Unlike traditional tools, ImageKit performs heavy-duty AI tasks—like background removal and neural upscaling—entirely within the user's browser using WebGPU and TensorFlow.js, ensuring data never leaves the local machine.

🌐 Live Demo

---

🚀 Key Features

1. Subject Architect (AI Background Remover)

• AI Extraction: Uses the RMBG-1.4 model via ImageKit WebComponents for instant, high-precision subject isolation.
• Manual Refinement: Integrated HTML5 Canvas tools for manual brush-based and box-based erasing using destination-out compositing.
• Custom Backdrops: Dynamic layering of solid colors or custom-uploaded images behind the extracted subject.

2. Resolution Architect (Neural Upscaler)

• Super-Resolution: Implements UpscalerJS to hallucinate missing pixels and upscale images by 2x, 3x, or 4x.
• Hardware Safeguards: Intelligent fallback to CPU if WebGPU/WebGL is unavailable, with patch-based processing to prevent system hangs.

3. PDF Deconstructor

• Frame Extraction: Converts multi-page PDFs into high-quality JPEGs using PDF.js.
• Selection Logic: Supports print-style page range selection (e.g., "1, 3, 5-8") for targeted extraction.

4. Filter Architect

• Cinematic Presets: Custom-engineered CSS filters for aesthetics like Polaroid, Retro, and Wong Kar Wai.
• Baking Engine: A "Canvas Baker" that captures real-time CSS filter styles and applies them directly to image pixels for high-quality export.

---

🛠️ Tech Stack

• Frontend: EJS (Embedded JavaScript Templates), CSS3 (Wireframe Aesthetic), Vanilla JS.
• Backend: Node.js, Express.
• AI/ML: TensorFlow.js, RMBG-1.4, UpscalerJS.
• Imaging: HTML5 Canvas API, PDF.js.

---

📦 Installation & Setup

1. Clone the repository:

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

2. Install dependencies:

   npm install

3. Start the local server:

   npm start

   The app will be available at http://localhost:3000.

🛡️ Privacy & Performance

This project is built on the principle of Edge Computing. By offloading image processing to the client's GPU:

• Zero Server Costs: The backend only serves static assets, making it infinitely scalable on free tiers like Koyeb.

• Total Privacy: Images are processed in RAM and never uploaded to any server, making it safe for sensitive documents.

🎨 Aesthetic

The UI follows a Retro-Wireframe design language, focusing on high-contrast borders and a clean, utilitarian "developer tool" feel.