PRACTICAL IMAGE EDITING

How to Remove a Watermark from a Photo for Free

For a watermark on an image you own or are allowed to edit, select the mark, protect nearby details, repair and compare before downloading. Start with a small area; complex or repeated marks may need more manual work.

Updated

Brush over the watermark

Example with a self-added translucent watermark
Actual Smart repair result after selecting the added watermark

Recorded Smart repair result

Illustrated brush selection using the actual area-tool mask and recorded Smart repair result. Playback speed is illustrative; results vary with the image and selection.

Try a translucent mark

Choose a small, clear selection

Mark in your photoPractical starting pointWhat to inspect
A corner word or logoBrush over the mark, including its edge or shadow.Nearby borders, gradients and remaining strokes.
A local translucent stampSelect its full visible footprint without covering unrelated details.Fabric texture or background patterns in the repaired patch.
Repeated text on several picturesFind text, use your own removal and keep words, then review each image.Missed lettering, incorrect matches and protected content.
A graphic repeated across much of the imageWork in small manual areas and assess whether a usable result is possible.Faces, seams or details hidden under the mark may not be reconstructed reliably.

Text recognition can suggest lettering; it is not automatic detection of every graphic watermark. Use the watermark editor for the brush workflow.

Brush, protect, repair and compare

  1. Open your own picture. Use JPG, PNG or still WebP, up to 10 MB, 16 megapixels and 4096 pixels per side. Opening a file does not start recognition or repair.
  2. Zoom into the mark. Brush is the starting tool. Paint short strokes over the watermark and any visible outline. Area adds a rectangle; Erase reduces the selection.
  3. Protect important details. Paint protection over a nearby face, edge, logo or fabric seam that must remain. A small selection is easier to inspect than a large sweep across the whole photo.
  4. Choose a repair. Try Quick repair on a small smooth area or Smart repair on texture. Smart repair may take longer on first use while required files load. Click the repair button only when the selection is ready.
  5. Compare before downloading. Inspect the repaired patch, nearby details and any remaining outline. Refine the selection and try again when needed. Download PNG or JPG before leaving the page.

The brush guide explains selection controls, while the background guide explains which export pixels can change.

Use word rules for repeated text marks

For a watermark made of recognizable text, choose automatic language detection or a supported language and use Find text. The first recognition asks permission to securely send a resized working image to TextWiper and its text recognition service providers. TextWiper does not save images or recognized text; providers handle data under their own policies. You can use the brush without recognition.

Enter your own removal phrase and optional keep words. Words is the default selection view, and Characters lets you refine smaller areas when real positions are available. Switching views preserves selections and does not scan again. Keep rules take priority; an empty removal list erases nothing.

A batch supports up to 10 images and 40 megapixels total. Review the matches, repair and download only successful results. Read the keyword guide and batch workflow. Graphic marks still need manual selection; text recognition has temporary usage limits.

Inspect the actual examples

The two examples below use original AI-generated photos with a corner logo and text, or three translucent SAMPLE marks we added ourselves. The recorded selections were made with the manual Area tool; the after-images are actual Smart repair PNG downloads from TextWiper. They are not the clean source photos used to create the fixtures.

In these examples, decoded pixels outside the actual removal mask, original alpha values, the face, the original SUMMER lettering and the image edges stayed unchanged. This describes the recorded examples, not a promise about every image. The corner edit looks smoother; the repeated translucent marks leave some local background softening. Inspect the repair before using it.

Common questions

Is the repair and download free?
Yes, during the current beta and within the stated image and batch limits. No account is required, and downloads do not add a TextWiper watermark. This does not promise unlimited processing or permanent pricing.
Will every watermark disappear automatically?
Choose the area yourself for graphic marks. Find text can suggest a text mark after upload confirmation, but it does not locate every logo, translucent pattern or repeated graphic.
Can I get the exact original pixels back?
Repair estimates covered content; it does not recover a guaranteed original. Fine texture can be softened or invented, and visible traces can remain. If essential information is covered, an unmarked source you are allowed to use may be preferable.
Which download format should I use?
PNG preserves transparency and decoded pixels outside the final removal selection. JPG flattens transparency to white and re-encodes the whole image. Neither format guarantees a perfect repaired patch.
What happens to my picture?
Brush selection, repair and downloads stay in your browser. Text recognition sends a resized working image only after the stated permission. Editing copies are held in page memory, and optional accounts do not save image histories. See the privacy details.

Worked examples: a corner mark and repeated translucent text

We added a corner symbol and SAMPLE lettering to one original AI-generated photo, and three translucent SAMPLE labels to another. The masks and results below were recorded with manual Area selections and actual Smart repair PNG downloads. The original SUMMER lettering on the shirt was kept.

Try this example in the editor

Loads only this site’s samples. No automatic scanning, removal or saved word rules. Follow the steps below.

  1. Load a marked example in the watermark editor. Loading selects no pixels and does not start recognition, repair or a download.
  2. Use Brush or Area to cover only the added marks. The recorded examples used two rectangles for the corner mark and three separate rectangles for the repeated labels, leaving the face and shirt lettering outside the removal mask.
  3. Choose Smart repair, explicitly run the repair, then compare the patch, nearby texture and image edges. The recorded result is from the editor, not the clean source photo.
  4. Download PNG after inspection. These exports kept decoded pixels outside the final removal mask and all original alpha values unchanged.

Corner symbol and text added to an original photo

Try this picture in the editor · Download this sample PNG

Three translucent SAMPLE marks added away from the face and shirt text

Try this picture in the editor · Download this sample PNG

White in the mask marks pixels allowed to be repaired. The dark background only makes the transparent mask visible.

The corner result is smoother, while the repeated translucent labels leave local background softening. These are two controlled examples, not a guarantee for every watermark. Repair estimates hidden content and may leave artifacts. The displayed masks are real Area selections; any animated brush path only illustrates how to cover their separate regions.

Original AI-generated photos with self-added marks · Real manual Area selections · Actual Smart repair downloads · October 12, 2026