The Gemini watermark is a small 4-pointed star logo placed in the bottom-right corner of every image generated by Google Gemini. It indicates the image was created by AI.

It comes in two sizes: 48×48 pixels with a 32px margin for images ≤1024px on either side, and 96×96 pixels with a 64px margin for larger images.

The logo is applied using alpha compositing — it is blended semi-transparently into the image pixels, not simply pasted on top.

“Nano Banana” is the nickname users gave to the Gemini AI star watermark. It originally referred to the Gemini Nano model family, but the term became shorthand for the 4-pointed star logo that appears on all Gemini-generated images.

You may see searches like “Nano Banana watermark remover” or “remove Nano Banana logo” — these all refer to the same Gemini star watermark that our tool removes.

Google adds the watermark as part of its AI transparency initiative. The visible logo signals to viewers that an image was AI-generated, not photographed or hand-crafted.

In addition to the visible star logo, Google also embeds an invisible SynthID watermark — a pixel-level provenance signal that persists even after the visible logo is removed.

Visible watermark: The 4-pointed star logo in the bottom-right corner. Semi-transparent, mathematically removable.

SynthID: An invisible watermark embedded directly into the pixel data by Google. It cannot be seen, does not affect image quality, and is not removed by any image editing tool including this one.

Our tool only removes the visible star logo. SynthID remains in the output image.

We use Reverse Alpha Blending — the mathematical inverse of how Gemini applied the watermark. The formula is:

original = (watermarked − α × 255) ÷ (1 − α)

Where α is the known transparency at each pixel and 255 is the white logo color. This recovers the exact original pixel values — not an estimate, not a guess.

Most other tools use AI inpainting that “hallucinates” what was underneath, often producing blur or incorrect textures. Our approach is mathematically exact, with an error margin of only ±1 pixel value — imperceptible to the human eye.

No. All processing happens entirely within your browser using the HTML5 Canvas API. Your image never leaves your device.

You can verify this yourself: open browser DevTools → Network tab, then process an image. You will see zero network requests made during processing.

The tool also works fully offline once the page has loaded.

Yes. The tool supports batch processing. Drop multiple images into the drop zone at once and they process automatically in sequence — no browser freeze, no extra steps.

For most images, the removal is pixel-perfect — the recovered area is mathematically identical to the original, with a maximum error of ±1 pixel value.

In rare cases where the image was resized, heavily compressed, or re-processed after Gemini generated it, the watermark pixels may not perfectly match the original alpha map, leading to very faint traces visible only at extreme zoom. This is a fundamental limitation of mathematical recovery, not a bug.

Input: JPG, JPEG, PNG, and WebP.

Output: Always lossless PNG for maximum quality. EXIF metadata is automatically stripped from the output for privacy.

No. Google applies the 4-pointed star watermark to all Gemini-generated images regardless of subscription tier. Gemini Advanced, Google One AI Premium, and Google Workspace plans all produce watermarked outputs.

The upgrade unlocks more powerful models, larger context windows, and premium features — not watermark-free images. Our free tool is the only way to remove the visible watermark after download.

No. The watermark is applied automatically at the system level — it cannot be disabled through prompts or settings. Every image Gemini generates will include the star logo, regardless of what you ask.

The only way to get a clean image is to remove the watermark after downloading, which is exactly what this tool does.

As of 2026, Google has not announced any plans to offer watermark-free Gemini images, even for paid subscribers. The watermarking policy is tied to Google’s AI content transparency commitments and the SynthID program.

It is possible this could change in future product updates, but there is no confirmed timeline.

Our tool itself is free for all uses. However, whether you can use the resulting images commercially depends on Google’s Terms of Service for Gemini and the laws in your jurisdiction.

For commercial use, review Google’s Generative AI Terms of Service to ensure compliance. When in doubt, consult a legal professional familiar with AI content licensing.

The most common use cases include:

Presentations & slides — Business decks, school presentations, and client deliverables where a third-party logo looks unprofessional.

Social media content — Instagram, X, LinkedIn, and TikTok posts where the star logo distracts from the visual story.

E-commerce & product images — Product mockups and listings on Shopify, Amazon, or personal websites.

Design mockups — Figma, Canva, and Photoshop projects where the watermark conflicts with the layout.

Blog & editorial images — WordPress or Ghost articles using AI-generated header images.

Yes. Images generated through Google AI Studio use the same Gemini watermark as images from the Gemini chat interface — the same 4-pointed star in the same position. Our tool removes it from both sources identically.