Question 1

Is this tool free?

Accepted Answer

Yes — completely free. No account, no watermark on results and no hidden limits.

Question 2

Are my images uploaded to a server?

Accepted Answer

No. The AI models run directly in your browser, so your images never leave your device.

Question 3

Which file formats are supported?

Accepted Answer

JPG, PNG and WEBP. You can also paste an image from your clipboard.

Question 4

What is the difference between MI-GAN and LaMa?

Accepted Answer

MI-GAN is fast and ideal for small watermarks and text. LaMa is a larger HD model that gives better results on large masked areas.

Question 5

Why does the first load take a while?

Accepted Answer

The models run in your browser and are downloaded once. MI-GAN is ~29.5 MB; LaMa (~208 MB) and the enhancer (~5 MB) load only when you use them.

Question 6

Which browsers are supported?

Accepted Answer

Chrome and Edge are recommended. A modern browser with WebAssembly (or WebGPU) support is required.

Question 7

Is this a real AI watermark remover?

Accepted Answer

Yes. The tool runs MI-GAN and LaMa deep-learning inpainting models directly in your browser. The AI analyzes the area around what you mark and reconstructs the missing pixels to match the surrounding image — no server processing, no upload.

Question 8

Can I also remove backgrounds and make ID photos here?

Accepted Answer

Yes — open the Cutout tab: the AI removes the portrait background automatically (soft hair edges preserved), then choose white, red, blue or a custom background and export the result.

Question 9

What size are the generated ID photos?

Accepted Answer

Presets at 300dpi: US / India passport 2×2 in = 600×600 px, UK / EU / Schengen 35×45 mm = 413×531 px, and Chinese 1-inch = 295×413 px / 2-inch = 413×579 px. The person is fitted to the frame automatically so the photo is ready to print.

Free AI Watermark Remover & Background Cutout

All tools