Remove Date Stamp from Photos

Remove the date stamp, time and "Shot on" text a camera printed on your photos. Finds it automatically, works in batches, runs in your browser.

Your files are processed in your browser and never uploaded.

Removal method

The AI model (about 17 MB) downloads once, then stays in your browser's cache.

Finds date stamps and time or place text printed near the top or bottom edge. Open each result to check it.

Drop images here

Choose files

HEIC, JPG, PNG, WebP, AVIF, GIF, BMP · up to 50 megapixels each · you can also paste

How to use

  1. Drop your photos or click "Choose files". You can add a whole folder's worth at once.
  2. Each photo is scanned for a date stamp or camera text near its top and bottom edges, and whatever is found is filled in to match the surroundings.
  3. Open a result to check it, then download the photos one by one or all together as a ZIP.

Features

Finds the stamp for you
No brushing or selecting. Orange date imprints, white time stamps, two or three lines of time, date and address, and "Shot on" captions are detected automatically.
Built for batches
Add hundreds of photos from an old camera or a phone that stamped every shot. They are processed in turn while you keep the tab open.
AI fill that continues the scene
An inpainting model rebuilds what was behind the text, so edges and textures carry on instead of turning into a smudge.
Private by design
Detection and filling happen on your device. Your photos are never uploaded; only the AI models themselves are downloaded, once.

Why date stamps end up on photos

Many cameras can burn the date and time into every picture. Older digital compacts and film cameras did it in orange or yellow digits in a corner, and a lot of Android phones offer a “watermark” setting that adds the time, date, place or the phone’s model along the bottom edge. The stamp is useful at the moment you take the picture, but it gets in the way when the photo goes into an album, a print, a presentation or a product listing.

Because the digits are part of the image itself, they cannot be switched off afterwards like metadata. Removing them means painting over the text with whatever should have been behind it.

How the detection works

With the AI method, a small text detection model (PP-OCRv5, about 2.5 MB) first marks every piece of text in a reduced copy of the photo. Each piece is then checked for what a camera stamp has and other text rarely does: a single flat ink that is pure white or orange, clearly brighter than what is around it, shaped like a line and close to the top or bottom edge. Text on screens, signs and whiteboards fails at least one of these tests and is left alone. In the accepted lines the letters are traced at full resolution together with the thin anti-aliased edge and drop shadow around them; leaving that halo behind would let the outlines of the digits show through after filling. The width of the halo is measured over the whole line rather than pixel by pixel, so a dark sleeve or a screen frame right next to the text is not mistaken for shadow and stays in the picture.

The Fast method needs no download and finds the text with rules alone: marks that are brighter, or more orange, than their surroundings, sit side by side on one baseline and share exactly the same colour are grouped into lines, and only lines near the top or bottom edge are accepted. It works well on plain backgrounds but tends to take in more of the surroundings.

How the text is filled

With the AI method, each stamp is cut out together with enough surrounding picture for context, scaled to the 512 × 512 pixels the model works with, filled, scaled back and blended in only where the text was. Long multi-line captions are processed in overlapping tiles so the letters do not shrink to a couple of pixels. The Fast method uses classic inpainting instead: thin strokes are filled from their edges inwards and thicker ones by copying matching patches from nearby.

Tips for the best result

Check a few photos from a batch before downloading all of them: tap a result to enlarge it. If a photo was not changed, its stamp is probably on a coloured box or away from the edge, which this tool does not handle. For a white border with a logo below the picture, use Crop Image instead, and if you want to strip location and camera details as well, run the results through Remove EXIF Data.

Frequently asked questions

Which stamps can it remove?

Text that a camera or phone printed onto the picture itself, near the top or bottom edge, in one solid colour. That covers the orange date imprint of older digital and film cameras, white time and date stamps, time-and-place captions from phone camera apps, and "Shot on" lines. If nothing like that is found, the photo is left exactly as it was and marked "No date or camera text found".

What can it not remove?

Text in the middle of the picture, dark text, captions on a coloured or semi-transparent box (common in timestamp camera apps), and frame-style captions where a white strip with a logo was added below the photo. A white strip is better cut off with the crop tool. This tool is meant for your own photos, not for removing other people's watermarks.

What is the difference between AI and Fast?

AI finds the text with PP-OCRv5, an open-source text detector, and fills it with MI-GAN, an open-source inpainting model that we trained further on camera stamps. Together they need a one-time download of about 17 MB, after which your browser keeps it in its cache. Fast needs no download, finds the text with simpler rules and works well for small text on sky, walls or other plain areas, but on busy backgrounds it can copy pieces of the letters back in.

Why does a processed photo look soft in one spot?

Very large digits that cross detailed structures, such as a screen edge or a patterned fabric, leave the model a big area to guess. It fills it smoothly, which can look slightly blurred at full size. Smaller stamps on ordinary backgrounds usually disappear without a visible trace.

Does it keep the photo's EXIF date and quality?

Processed photos are saved again as JPG (HEIC photos too), PNG, WebP or AVIF at high quality, and like most editors the tool does not copy EXIF data such as the date taken into the new file. Photos where nothing was found are returned untouched, metadata included.

How do I stop my camera from adding a date stamp?

On many Android phones, open the Camera app's settings and turn off "Watermark", "Time watermark" or "Shot on" (the name depends on the brand). On compact and film cameras the option is usually called "Date imprint" or "Date print" in the setup menu. iPhone's own Camera app never prints a date on photos; third-party timestamp apps do.

Related tools