Google AI Studio
Watermark Remover
Images generated in Google AI Studio carry the exact same Nano Banana star watermark as Gemini. Remove it instantly — free, browser-based, completely private. No file ever leaves your device.
Is the Google AI Studio Watermark the Same as the Gemini Watermark?
Yes — completely identical. Whether you generate an image on gemini.google.com or inside aistudio.google.com, Google applies the same Nano Banana star watermark using the same fixed process. The position is always the bottom-right corner. The opacity map is the same. The two size variants — 48×48 pixels for smaller images and 96×96 pixels for larger outputs — are identical across both platforms.
This matters because any tool that correctly removes the Gemini watermark handles AI Studio images without modification. There is no separate AI Studio watermark, no different alpha map, no platform-specific variant. The underlying image generation model — Google's Imagen / Nano Banana — is shared, and so is the watermark it applies.
Where the two platforms differ is not in the watermark itself but in how you save your images before removal. That difference directly affects the quality of the result.
How Google AI Studio Saves Images Differently — and Why It Matters
On gemini.google.com, saving an image is simple: right-click and choose Save Image As. The file you get is the original PNG straight from Google, with pixel values untouched.
Inside Google AI Studio, images appear inside the prompt response panel. The correct way to save is using the download button — the downward arrow icon that appears on hover within the generation card. This delivers a clean, uncompressed PNG identical in quality to what Gemini produces.
The problem is that many AI Studio users take a screenshot instead of using the download button. Screenshots on Windows, macOS, iOS, and Android all apply compression and colour profile conversion that permanently alters the pixel values in the watermark region. The reverse alpha compositing formula works by recovering exact original pixel values — once a screenshot has changed those values, the formula produces a faint colour fringe instead of a clean result.
The same applies to images shared through Slack, WhatsApp, or Telegram before processing. Each transfer that re-encodes the file degrades the input. Always process the original downloaded file.
Always use the AI Studio download button, not a screenshot. The downward arrow icon in the generation card saves the original file. Screenshotting — even at full resolution — introduces compression that breaks the reverse formula and is the most common cause of imperfect results.
Three Ways to Remove the Google AI Studio Watermark
AI Studio users range from designers testing prompts one image at a time to developers running generation pipelines at scale. The right method depends on your workflow.
Download your image from AI Studio using the download button, then upload it to the QuickImageFix tool. It auto-detects the watermark size, applies the reverse formula in your browser, and returns a pixel-perfect result in under 100ms. No account, no server upload, batch up to 10 images at once. Works on every browser — desktop, iPhone, Android.
If Google AI Studio is part of your regular workflow, the Chrome extension removes the watermark automatically at the point of download. It intercepts the download event on aistudio.google.com, applies the removal, and delivers a clean file to your downloads folder without any manual step. Install once and never think about it again. View the extension installation guide →
If you are generating images programmatically via the Gemini API and need watermark removal integrated into the pipeline, the underlying reverse alpha compositing algorithm is fully open-source. Allen Kuo's original implementation and community Python forks are available on GitHub with the alpha map files for both watermark size variants. A full Python guide is coming soon on this site.
Step-by-Step: Removing the Watermark from an AI Studio Image
The process is straightforward. The only AI Studio–specific step is how you save the image before uploading.
Open aistudio.google.com and generate your image as you normally would. Nothing in the generation workflow changes.
Hover over the generated image in the response panel. Click the downward arrow download icon to save the original PNG. Do not screenshot the output — screenshots permanently alter the pixel values that the removal formula depends on.
Open the Google AI Studio watermark remover online tool. Drag and drop your saved PNG — or tap to upload on mobile. You can batch process up to 10 images at once.
The tool detects the correct watermark size variant automatically and applies the reverse formula locally in your browser in under 100ms. Download as PNG, JPEG, or WebP — or grab all results as a single ZIP file.
Necessary only if you are posting to Instagram, LinkedIn, or Pinterest. Removing the visible star is not enough — those platforms read an invisible C2PA credential and still show "Made with AI." Run the cleaned image through the AI Metadata Scrubber to remove that credential as well.
Ready to process your AI Studio images? The tool runs entirely in your browser — no account, no upload to any server.
Does Removing the Watermark Also Fix the "Made with AI" Label?
No — and this is the most common point of confusion for AI Studio users who publish images to social platforms. Removing the visible Nano Banana star solves the obvious problem. But Google AI Studio images carry a second, completely separate invisible layer: a C2PA content credential embedded in the file's metadata.
C2PA is a cryptographic standard that records how an image was created — which model generated it, when, and through which platform. Instagram, LinkedIn, and Pinterest read this credential automatically and display a "Made with AI" label on any image that carries it, regardless of whether the visible watermark has been removed.
For most use cases — presentations, documents, design mockups — the C2PA credential is irrelevant. But if you are publishing to social platforms, stripping it is a separate step. The online AI Metadata remover removes C2PA tags and AI-origin markers without touching your EXIF camera data.
| Layer | What it is | Status |
|---|---|---|
| Visible Nano Banana star | The four-point star overlay in the bottom-right corner, applied via alpha compositing | Removed by the QuickImageFix tool |
| C2PA content credential | Invisible cryptographic metadata tag — triggers "Made with AI" on Instagram, LinkedIn, Pinterest | Requires the AI Metadata Scrubber |
| SynthID frequency signal | Imperceptible signal embedded in frequency data by Google DeepMind — does not affect visual appearance | Cannot be removed by any current tool |
Can You Generate Images in AI Studio Without the Watermark?
No. Google applies the Nano Banana watermark to all images generated through the Gemini image model regardless of which interface you use — gemini.google.com, aistudio.google.com, or directly through the Gemini API. There is no setting, no plan tier, no API parameter, and no prompt technique that disables the watermark at generation time.
The watermark is applied server-side during generation before the image is returned to you. Google's stated rationale is AI image identification for content transparency purposes — the visible star is part of the same system as SynthID. Both are applied automatically and cannot be suppressed from the user side.
This means removal must happen client-side after the image has been downloaded. For the technical explanation of how the reverse formula works, see How Gemini Adds Watermarks to Its Images and How to Remove Them Without Losing a Single Pixel.