Summary

This project provides a privacy-focused Chrome extension to remove watermarks from Google Gemini-generated images. It offers a local, standalone solution, ensuring user data never leaves the browser. This benefits users prioritizing data security, image integrity, and a seamless processing experience without server reliance.

How It Works

The extension leverages the LaMa (Large Mask Inpainting) AI model, executed client-side via ONNX Runtime Web. This architecture enables high-quality watermark removal through intelligent inpainting, preserving the original image's pixel-perfect quality by modifying only the masked watermark area. All processing occurs locally within a sandboxed browser environment, guaranteeing complete user privacy and data security.

Quick Start & Requirements

• Installation: Manual installation required as an unpacked Chrome extension. Detailed, step-by-step instructions are available in the INSTALLATION.md file.
• Prerequisites: Google Chrome browser.
• Usage: Load the extension, drag-and-drop a Gemini image into the interface, and download the processed output using the provided comparison slider.

Highlighted Details

• 100% Local & Private: All image processing and AI inference occur directly within the user's browser, ensuring no data transmission.
• LaMa Inpainting: Utilizes a powerful, resolution-robust AI model for high-fidelity watermark removal.
• Quality Preservation: Modifies only the small watermark area, leaving the rest of the image untouched and maintaining original quality.
• Modern & Intuitive UI: Features drag-and-drop support, a real-time progress bar, detailed logs, and an interactive before/after comparison slider.
• Efficient Performance: The AI model is loaded once and cached for subsequent uses, supported by a modular JavaScript architecture.
• Open Source: Fully transparent codebase encouraging community contributions.

Maintenance & Community

Contributions are welcomed via issues and pull requests, with detailed guidelines provided in the CONTRIBUTING.md file. No specific community channels (e.g., Discord, Slack) or maintainer details are listed in the README.

Licensing & Compatibility

• License: Apache License 2.0.
• Compatibility: The permissive Apache 2.0 license makes it suitable for commercial use and integration with closed-source applications without significant restrictions.

Limitations & Caveats

Requires manual installation as an unpacked Chrome extension. Functionality is specifically targeted at removing watermarks from Google Gemini-generated images. Performance is dependent on the user's local browser and hardware capabilities for running the ONNX model.