A powerful, AI-driven tool designed to automatically remove the Google Gemini "Sparkle" watermark from images while preserving high quality.
(No installation required. Test it now!)
See the magic yourself.
| Before (With Watermark) | After (Cleaned by AI) |
|---|---|
![]() |
![]() |
![]() |
![]() |
The easiest way to run the application without manual setup.
Simply double-click start.bat.
Open your terminal and run:
bash start.shThe application will automatically install dependencies and launch at http://localhost:8000.
If you prefer to run it manually or modify the code:
- Python 3.10 or 3.11
- PyTorch (CPU or CUDA)
# 1. Create Virtual Environment
python -m venv venv
source venv/bin/activate # Windows: venv\Scripts\activate
# 2. Install Dependencies
pip install -r requirements.txt
# 3. Run Server
python server.pyRun the optimized container (CPU-friendly, ~3GB) without valid Python setup.
- Start: Run
docker-run.sh(Mac/Linux) ordocker-run.bat(Windows). - Image:
akashyadav758/nanoremover:latest
A quick overview of the codebase:
NanoRemover/
├── server.py # Main FastAPI Backend
├── requirements.txt # Python Dependencies
├── start.sh / .bat # Auto-start automated scripts
├── docker-run.sh / .bat # Docker launch scripts
│
├── src/
│ └── model_handler.py # LaMa AI Model Logic
│
└── ui/ # Frontend Resources
├── index.html # Main Interface
├── style.css # Glassmorphism Styling
└── app.js # Logic (Uploads, Previews)- Gemini Watermark Removal: Specifically trained to target the watermark position.
- Smart UI: Glassmorphism design with Dark Mode.
- Instant Preview: Generates fast WebP previews for UI.
- Privacy Focused: Auto-deletes uploads after 1 hour.
- Cross-Platform: Works on Windows, Mac (Intel/Silicon), and Linux.
Made with ❤️ by Akash Yadav



