---
license: mit
tags:
- image-segmentation
- background-removal
- salient-object-detection
- image-processing
- onnx
- webgpu
- webassembly
---

RemoveG is a browser-based AI image background remover designed to process images locally on the user's device.

🌐 Website

https://www.removeg.io/

✨ Features
Automatic image background removal
Client-side image processing
WebAssembly and WebGPU acceleration
Support for PNG, JPG, JPEG, and WebP images
Designed for fast browser-based inference
No account required for basic usage
Privacy-focused image processing
🧠 Model

RemoveG uses the IS-Net / InSPyReNet approach for salient object detection and foreground extraction.

The model can be useful for:

Product photography
Portrait images
E-commerce images
Logos and graphics
Signatures
General foreground extraction
⚡ Browser Inference

The RemoveG web application uses ONNX Runtime Web to run machine-learning inference directly in the browser.

Depending on browser and hardware support, inference can utilize:

WebGPU
WebAssembly
Multi-threaded WebAssembly

This approach allows image processing without requiring a traditional remote image-processing API.

🔒 Privacy

RemoveG is designed with privacy in mind. Image processing is performed on the user's device through browser-based inference.

Learn more at:

https://www.removeg.io/

🛠️ Technology
Component	Technology
Background Removal	IS-Net / InSPyReNet
Model Runtime	ONNX Runtime Web
GPU Acceleration	WebGPU
CPU Runtime	WebAssembly
Processing	Client-side
Output Formats	PNG, JPG, JPEG, WebP
📌 Use Cases
E-commerce

Remove backgrounds from product images to create clean product photography.

Portraits

Create transparent-background profile pictures, headshots, and other portraits.

Graphic Design

Extract foreground objects for thumbnails, advertisements, presentations, and social media graphics.

🔗 Learn More

Visit RemoveG.io to try the browser-based image background remover.

⚠️ Model Information

This project uses the IS-Net / InSPyReNet approach for image salient-object detection. Please refer to the original model implementation and its associated license for details about the underlying model.

📄 Keywords

background-removal image-segmentation IS-Net InSPyReNet ONNX ONNX-Runtime-Web WebGPU WebAssembly image-processing foreground-extraction browser-AI