toolgarden.xyz
中文
watermark removalAI inpaintingimage repairlocal processing

Why Does Image Watermark Removal Leave Blur or Ghosts? How Local AI Inpainting Works, and Its Limits

Removing a watermark is not erasing it; a model repaints the area from the surrounding pixels. This guide uses concrete numbers to show how selection size drives the result, how to pick among the three modes, and which watermarks you should not hand to AI inpainting at all.

ToolGarden tools prioritize browser-local processing, so files and text do not need to be uploaded to a server.

Published July 2, 2026Updated October 14, 202612 min readBy ToolGarden

Almost everyone who tries AI watermark removal hits the same surprise: you select the area and click generate, the watermark is gone, but where it used to be there is a smeared patch of color or a faint shadow. The tool is not broken. There is a gap between what you expect from “removing” a watermark and what actually happens: it does not restore anything, it repaints.

This guide first covers which images are fair game, then explains how ToolGarden repairs an area inside the browser, then walks through two numeric examples showing how selection size decides the outcome, and finally lists how to diagnose ghosts, soft patches and seams, plus the cases where you should not use this approach. The AI models run locally in your browser, and the image is not uploaded to a server.

1. Draw the line first: only edit images you have the right to modify

A watermark usually stands for a copyright or source claim. Removing it from someone else’s work to republish, sell or pass it off as your own can infringe copyright and breaks the terms of most platforms. The tool is meant for images you are entitled to process, for example:

  • Your own photos with a date-time stamp burned into a corner by the camera.
  • Assets you made and watermarked yourself, when only the watermarked copy survives.
  • Your own promotional images that still carry an old logo after a rebrand.
  • Screenshots where you want to remove an account name, a temporary label or personal information.

If the image comes from someone else, get permission first or ask the rights holder for the unwatermarked original. The tool’s own page carries the same reminder: only process images you own or are allowed to modify.

2. How it works: repair is not restoration

The pixels under a watermark have been overwritten, so the original information is not in the file, and no tool can restore it. All it can do is guess what should be there from the surrounding picture and paint that in. The technique is called inpainting. ToolGarden’s pipeline is:

  1. You drag a rectangle on the image; that rectangle is the region to repaint (the mask). The mask is currently a rectangle, not a brush stroke.
  2. The tool crops a patch of the image that includes context around the selection and passes it to the model.
  3. The model repaints only the pixels inside the mask; pixels outside the selection stay unchanged.
  4. The result gets one more round of edge softening (the “Edge blend” control) to reduce the seam.
ModeHow it worksBest forCost
High quality AI (default)The MI-GAN ONNX model, run locally in the browserPhotos, textured or gradient backgroundsThe first run downloads about 28 MB of model assets
Compatible AIThe LaMa ONNX model, run locally in the browserPhotos and richer textures, as a compatible fallbackAlso needs a model download
Fast localNo AI model; fills by distance-weighted estimation from nearby pixelsFlat colors, gradients, screenshot backgroundsSmeared look on complex texture

There is also automatic degradation between the modes: if you pick High quality AI and the model fails to load or run, the tool tries LaMa, and if that fails too it drops to fast local repair, with a progress message saying it is switching to fast local repair. So after generating, glance at the Mode field in the result details to confirm which one produced the output.

3. The procedure

  1. Upload an image. JPG, PNG, WebP, GIF, BMP, SVG and AVIF are accepted, up to 50 MB and 40 million pixels.
  2. Drag the selection. Once an image loads, a default box is placed at the bottom right: 32% of the image width, 14% of the height, with 5% margins from the right and bottom edges. On a 1920×1080 image that is 614×151 px with its top-left corner at (1210, 875). Drag it, resize it with the eight handles, or type values into the X, Y, Width and Height fields.
  3. Pick a mode. High quality AI is usually the first choice; on a flat color or smooth gradient, Fast local is far quicker.
  4. Adjust Edge blend. The range is 0 to 36 px, default 12. It grows the mask a few pixels outward and softens the seam.
  5. Choose the output format. The default follows the source: JPG gives JPG, WebP gives WebP, anything else gives PNG. JPG and WebP also have a quality slider from 50% to 100%, default 92%.
  6. Click Generate repair, zoom in on the preview, and download when satisfied. Moving the box or changing parameters clears the previous result, so regenerate afterwards.

4. Worked examples: how selection size decides the outcome

Example 1: a 1920×1080 screenshot with a text watermark at the bottom right

Say the watermark text occupies 220×40 px with its top-left corner at (1650, 1000). Do not box it exactly; leave about 10 px on every side: x=1640, y=990, width 240, height 60. The reason is that watermark edges contain anti-aliased semi-transparent pixels, and sometimes a stroke or shadow, so a box that is too tight leaves a faint outline of the letters. If the background is a flat color or smooth gradient, Fast local finishes almost instantly; on a textured background use High quality AI.

Example 2: a 4000×3000 photo, default box versus a tight box

This is the key to understanding result quality. The AI model input is a fixed 512×512 pixels, so the tool crops a patch with context around the selection and shrinks it to 512 before handing it over. Reading the source, the context size is the largest of three numbers: 48 px, 8% of the shorter image side, and 80% of the selection’s longer side; the crop is a square whose side is the selection’s longer side plus twice the context size, capped at the shorter image side.

SelectionContextSquare crop sideScale to reach 512
Default 1280×420max(48, 240, 1024) = 10243000 (capped by the short side)Shrunk to about 17% of original
Tight 500×140 around the logomax(48, 240, 400) = 4001300Shrunk to about 39% of original

In other words, on the same 4000×3000 photo the default large box has the model painting on a version compressed to roughly one sixth, which is then scaled back up, so the repaired area ends up visibly softer than the grain around it. With a tight 500×140 box over the actual logo the model sees a much sharper picture. The takeaway is plain: the closer the selection hugs the watermark, the better, and for a large watermark it is better to work in several small passes, reloading the previous PNG each time, than to box a huge region at once.

5. Choosing an approach per watermark type

Watermark situationRecommendationExpected result
Semi-transparent corner logo on a textured photoHigh quality AI; box hugging the logo with a 5 to 10 px marginUsually natural
Text on a flat or gradient backgroundFast local is enoughFast and hard to spot
Camera date stamp (orange digits)Small box, High quality AIGood on simple backgrounds
Repeated watermark tiled across the whole imageNot suitableThe selection covers nearly everything, leaving the model nothing to reference
Watermark over faces, text or product detailNot suitableHidden detail cannot be recovered; you get plausible invented content
Watermark over regular patterns such as brick, grids or tidy textBe careful; try several small passesRegular textures tend to misalign or repeat

6. Common problems and how to diagnose them

SymptomMost likely causeWhat to do
A faint outline remainsBox too tight; anti-aliased edges and shadow not coveredGrow the box a few pixels, or raise Edge blend to 18 to 24
The repaired area is blurry or softSelection too large relative to the image; model input shrunk heavilyShrink the selection; handle large watermarks in several passes
Color differs from the surroundingsSelection crosses a clear color boundaryKeep each selection inside one color region and do them separately
Odd repeated texture appearsToo little context to reference, or the box is too largeShrink the box; try the other AI mode
Slight blocky noise in the resultOutput as JPG at low quality, or compressed repeatedlyOutput PNG for intermediate results and export JPG only at the end
The first run is very slowAbout 28 MB of model data is downloadingWait for it; later runs in the session do not repeat it
Mode in the result says Fast localModel loading failed and the tool fell backCheck the network, refresh, or choose Compatible AI

Metadata and quality

The output is a new file re-encoded through a Canvas, so the original EXIF data (capture time, device, location) is not kept. If you need to know what metadata the original carried, look at it in the EXIF viewer first. Also, outside the selection, pixels are unchanged when you export PNG; exporting JPG or WebP re-compresses the whole image, producing slight overall changes, so choose PNG when you need the highest fidelity.

Choosing the export format

FormatWhen to use itWatch out for
PNGFurther editing, intermediate results, transparent areasLargest files; photos grow noticeably
JPGFinal use on the web, social platforms or printLossy; transparent areas are filled with white
WebPWeb delivery with smaller filesSome older software cannot open it

7. When not to use this approach

  • Images you have no right to modify: a watermark is a rights statement, and republishing after removing it may be infringement.
  • Images that must be kept as evidence or records: the repaired region is model-generated content and can no longer count as the original record.
  • Watermarks that cover most of the frame or run across the whole image: there is not enough surrounding information, and you should ask the rights holder for a clean version.
  • Invisible digital watermarks: this tool handles visible watermarks only and does nothing about hidden marks embedded in image data.
  • Needing to recover obscured text or faces exactly: AI generates plausible content; it cannot recover the real information.

Frequently asked questions

Q.Why is the repaired area blurry?

The most common reason is a selection that is too large relative to the image. The model input is a fixed 512×512, so a big selection pulls in a big context that is then shrunk heavily and yields a low-resolution fill. Shrink the selection to hug the watermark, or work in several small passes, and the result usually improves clearly.

Q.How do I choose between the three modes?

Use the default High quality AI for photos and textured backgrounds; switch to Compatible AI if it misbehaves; choose Fast local for flat colors, gradients and screenshots, where it is quick and needs no model download. If unsure, try Fast local first and move to AI if the result is unsatisfying.

Q.Is my image uploaded to a server?

No. The models run in your browser and the image is processed in the current tab’s memory. The first run downloads model files, which is a download and carries none of your image; data is cleared when you close the page.

Q.Can I remove watermarks in batch?

Currently the tool handles one image and one rectangular selection at a time. For several watermark positions, work in passes: generate and download the result, load it again, and select the next position.

Q.Why does the output lack capture time and other information?

The output is a new image re-encoded through a Canvas, and EXIF metadata is not written into it. That helps privacy, but if you need the shooting data, record it separately.

Q.Can I remove watermarks from other people’s images?

It is not advisable, and you should not. A watermark usually states copyright or origin, and removing it without permission and republishing can be infringement. Get authorization, or ask the rights holder for an unwatermarked original.