How to Remove Text, Captions and Date Stamps from a Photo: Online, on a Phone and in Photoshop

The simplest way to remove text from a photo is to paint over the letters with a brush and let a neural network fill the gap, rebuilding the background from the areas around it. On sky, a wall, paper or asphalt the repair is usually invisible. On a face, a pattern or fine detail there will be no miracle, and it is better to know that in advance so you do not waste your attempts.

There are four parts below: which text comes off cleanly and how to do it online, what to do about a date and time in the corner of the frame, how to work from a phone, and how to remove text by hand in Photoshop. All the examples are about your own pictures: a camera date, handwriting on a postcard, a random shop sign behind you, an extra line on your own screenshot or layout.

Removing text from a photo: what the result depends on

The result depends not on the text itself but on what surrounds it. The program does not rub the letters out the way an eraser rubs out pencil. It cuts out the whole area and fills it in again, looking at the edges. The simpler and more uniform the surroundings, the more accurate the patch.

So before you upload a picture, look at the background under the text and judge three things: how uniform it is, how much room the text takes up, and whether any sharp lines run through it (the horizon, the edge of a table, a tile joint).

Text on a plain background: sky, wall, paper

This is the most rewarding case: text on sky, on a painted wall, on a sheet of paper, on sand or water goes away in a single pass. The background repeats itself there, and all the neural network has to do is carry the colour and grain from the edges inwards.

These come off well:

• date digits in a dark corner or on the sky;
• a pen inscription in the margin of a postcard or an old photograph;
• a line of text on a solid colour block in your own layout;
• a small sign or plaque on a flat wall in the distance;
• a number or a note on the light backing of a scan.

A smooth change of tone (a sunset sky, a shadow on a wall) is not a problem either, as long as the text is small. Large text on a gradient sometimes leaves a blotch of a slightly different shade that shows when you zoom in.

Text over a face, a pattern or fine detail

The honest answer here is that it will not come out clean, and the more important the detail under the letters, the more visible the mark. If the text runs across a face, the neural network will draw skin in its place, but it will not bring back the real eye, lips or mole. You get a plausible patch, not a portrait.

The same goes for any regular pattern: brickwork, a checked shirt, fence railings, wallpaper, the knit of a sweater. Inside the patch the pattern usually loses its rhythm and the lines drift a little. Small print under the caption (for example, a line written across a book page) turns into unreadable ripple.

What you can do in these cases:

• paint over only the letters themselves with a thin brush instead of covering the whole line with a rectangle: the smaller the area, the less the program has to invent;
• crop the frame if the text sits near the edge;
• finish the job by hand in an image editor, which is what part four is about.

Removing text from an image without damaging the background: what that really means

Taken literally, you cannot remove text without damaging the background, because there is no background under the letters: at that spot the file holds only the pixels of the letters themselves. No program knows what used to be beneath them. It builds something that looks right next to the neighbouring areas.

Two practical conclusions follow. First: on a simple background the rebuilt part cannot be told from the real thing, and to the eye the background really is intact. Second: on a complex background what you get is a guess. If the text was covering something important (a house number, a face, a detail of a pattern), it will not be restored. This also applies to text that has been painted over or scribbled out: the tool will not help you read it back.

AI text remover: how the online tool works

In the Remove objects from photo tool you show for yourself what should be erased: you paint over the text with a brush and the neural network fills in whatever you painted. There is no automatic text detection, and that is more of an advantage: the program will not touch a sign you wanted to keep.

What you should know before uploading:

• it accepts JPG, PNG and WebP files up to 25 MB, with sides from 100 to 8000 px; convert HEIC from an iPhone to JPG first;
• one file is processed at a time;
• the result comes back as a JPG; if the long side is over 4096 px, the output is scaled down to that size;
• no registration is needed.

Processing happens on the server, so it puts no load on your phone or an old laptop. All you see on screen is a spinning indicator, and after it the finished picture appears.

How to erase text from a photo step by step

The whole job takes just a few actions:

1. Upload the picture: drag the file into the box or choose it from your device.
2. Use the Brush slider to set the thickness: from 6 to 120, measured in pixels of the picture itself. For a line of text it is handy when the brush is slightly taller than the letters.
3. Paint over the whole of the text, with a small margin around the edges. A translucent red fill shows what will be removed.
4. If you have gone over something you need, switch on the Eraser and wipe off the extra fill. The Clear mask button removes everything you painted in one go.
5. Press Remove and wait for the result.
6. Check the preview. If you do not like it, adjust the fill and run it again, or press New photo. The fill stays in place after processing, but every run counts as a separate free edit.

If you press Remove without painting anything, the tool will ask you to mark an area with the brush first.

Why you should paint with a margin

You need a margin because letters almost always have a halo: a shadow, an outline, a blurred edge left by JPEG compression. If you paint strictly along the contour, the halo stays, and a ghost of the text can still be read where it used to be.

Going a few pixels past the edge of the letters is enough. A fill that is too wide does harm as well: the larger the area, the more the neural network has to make up. The happy medium is to cover the letters and their shadow without catching neighbouring objects. If a sharp boundary runs next to the text (the edge of a frame, the horizon line), try not to touch it, and it will stay straight.

A long caption: remove it all at once or in pieces

In the online tool it is more convenient to paint over the whole caption at once: in a single run you can mark as many areas as you like in different parts of the frame. Each run works on the file you uploaded, so the results of separate runs do not add up by themselves.

Trying it piece by piece does not pay off here: you get 5 free edits a day, and every test on a single word uses one of them. It is better to paint everything carefully straight away. If the line crosses different surfaces, say half on the sky and half on a roof, change the brush thickness as you go: on the sky you can paint more broadly, on the roof keep strictly to the letters.

Letters taller than 120 px on a large picture cannot be covered with one stroke: run the brush in two or three overlapping rows, leaving no gaps between them.

Removing text from a photo online without losing quality: the honest answer

No editor works with absolutely no loss, but the loss here is small and easy to understand. Anything you did not paint over is not redrawn: on large pictures the tool takes the original image and replaces only the marked zone in it, with a soft edge. The file is still re-saved as a JPG, as with any editing.

Two details matter for big pictures. A frame with a long side over 4096 px will be scaled down to that size. And the patch itself, on a photo larger than 1536 px on the long side, may look slightly softer than its surroundings, because the neural network draws it on a copy reduced to that size. On sky and walls you will not see it, on a crisp texture up close you might.

A date on a photograph comes in two kinds, and they are removed in completely different ways. The first is digits printed right into the picture, usually yellow or orange, in a corner. The second is the capture date written into the file properties, which is not visible on the image itself. Work out which one you have first, otherwise you can spend a long time looking for the wrong button.

Removing a date stamp from a photo: what those digits in the corner are

The digits in the corner of the frame are put there by the camera itself at the moment of shooting, if date imprinting is switched on in its menu. The feature comes from film point-and-shoot cameras and is still found in compact cameras and in the camera apps of many phones.

The main thing to understand is that a date stamp does not sit on top of the picture as a separate layer. It becomes part of the image, pixels just like the sky and the grass. Canon's help pages for its compact cameras say plainly that a date stamp cannot be edited or removed after shooting, and Samsung says the same about the watermark that the Galaxy camera adds. There is no button that lifts the date off a finished shot: the digits can only be painted over and the background rebuilt.

How to remove the date from a photo on a computer

On a computer the date is removed like any other text: you paint over the digits and the program redraws the corner. In that sense a date stamp is an easy patient: it is small, it sits near the edge, and most often it falls on a calm background such as the floor, grass, sky or the dark bottom of the frame.

1. Open the picture in the online tool in any browser.
2. Make the brush smaller with the slider so that it is slightly taller than the digits.
3. Paint over the date in one stroke with a small margin: bright digits often have a coloured halo.
4. Press Remove and check the corner in the preview.

If the date has landed on someone's legs, the edge of furniture or a patterned carpet, the result will depend on the pattern. In simple cases one attempt is enough, in difficult ones it is easier to finish the corner by hand or crop it off.

Remove unwanted objects from a photo

Brush over the object that gets in the way (a watermark, a person in the frame, wires) and AI rebuilds the background. Local, no subscription.

Drag a photo here or click to choose JPG, PNG, WebP. Up to 25 MB.

Brush 40 Brush Eraser Clear mask

Paint over the object fully, you can go a bit beyond the edges. The eraser fixes any excess.

Remove New photo

AI is removing the object...

Done!

A hard case? Manual retouching in Photoshop

AI is great on simple backgrounds. If there is an important scene behind the object (faces, fine detail, a continuation of the object), a human hand delivers a result with no AI traces. From $10 per photo.

Order manual retouching

×

Open the full tool page →

Removing the date from a photo online for free: how many tries and what you get

Processing and the preview are free: 5 edits a day are available from one address, and that count is shared by all the tools on the site that run on neural networks. The first edit of the day starts at once, before the following free ones you will have to wait in a queue for about two minutes, with a timer shown on the screen. Next to the timer there is a paid Skip the queue button: the picture is processed immediately, but in that case there is no preview, and the finished file opens after payment.

The free preview is reduced to 900 px on the long side and carries the site's mark over the picture. That is enough to judge whether the date came off cleanly. The clean full-size file is paid. A sensible order is this: first make sure from the preview that the corner looks fine, and only then decide whether you need the full file.

When cropping the date is easier than erasing it

If the digits sit in the very corner and there is nothing important nearby, cropping gives a result without a single invented pixel. Trim a narrow strip off the bottom or the side, and the date is gone.

Cropping works when:

• there is spare room at the edges of the frame and the composition will not suffer;
• the picture does not need to be printed in a strict format where the proportions matter;
• you have many such photographs and want to treat them all the same way.

It does not work if the date runs onto people, if the frame is already shot tight, or if it is a scan of a paper print that you want to keep whole. Then painting over is what remains.

How to turn off the date stamp on a digital camera

On the camera itself you cannot remove the date from a shot that has already been taken, you can only switch it off for future pictures. On compact cameras an item in the shooting menu is responsible for this. In Canon's help for the compact PowerShot and IXUS models it is called Date Stamp, and it has three positions: Off, Date and Date & Time, with the digits placed in the lower right corner of the frame. In other menu languages and on other brands the name differs, so look for a line with the word “date” in it and check the manual for your model.

The sequence is the same everywhere: the MENU button, the tab with shooting settings, the line about the date, the off position, MENU again. Do not confuse it with the camera's clock setting (Date/Time): that one sets the time for the file properties and prints nothing on the image.

For shots already taken there is one route: move the files to a computer or phone and paint over the digits as described above.

How to remove the date from photos on a Samsung

On a Samsung Galaxy the date is added by the camera, not by the gallery, and it is switched off in the camera settings. According to Samsung's help the order is this: open the Camera app, tap the settings icon and find the item that adds a watermark with the model name to the picture. Inside it there are separate lines for the model name and for the Date and time option. You can turn off the date alone or the whole item.

The setting applies to new shots. In the same help page Samsung warns that such a watermark cannot be taken off a finished photograph. Old pictures with a date will have to be fixed as images: paint over the digits in the online tool or use object removal in Gallery, which is covered in the next part.

How to remove the date from photos on Xiaomi and other Android phones

On Xiaomi, Redmi and POCO the principle is the same: the time stamp is switched on in the settings of the camera app, and that is where it is switched off too. Open the camera, go into its settings and find the section that deals with watermarks on pictures, with a toggle for the time stamp. The name of the section changes from one version of the interface to the next, so go by the meaning, not by the exact words.

On other Android phones (Honor, Huawei, realme, Tecno and others) the route is similar: camera, settings icon, the item about a watermark or a stamp on the photo. If there is nothing of the kind in the camera settings and the date still appears, check whether you are shooting with a third-party camera app: it has settings of its own.

How to remove the time and date from a photo on Android when the shot is already taken

From a finished picture the time and date are removed only by editing the image. Turning off the setting in the camera has no effect on old shots, however many times you toggle it.

The options, from least to most effort: crop the edge in the gallery, paint over the digits in the online tool through the phone's browser, or use the object removal built into the gallery, if your model has it. For a dozen family photos the first or second way is the quickest.

How to remove the date from a photo's file properties

The date in the file properties, that is, in the EXIF metadata, has nothing to do with the picture: it is a service record inside the file where the camera stores the capture time, the device model and sometimes the coordinates as well. No brush or neural network is needed here, and the image does not change at all.

On a Windows computer this information is shown in the file properties on the Details tab, and it can be cleared there too. It is quicker to do it online: the EXIF cleaner tool strips the metadata from the file completely. What exactly is stored in a picture's properties and what the risks are when you send it on is explained in the article on how to remove EXIF data and location from a photo.

The reverse is true as well: cleaning the metadata will not remove digits printed into the corner of the frame. If you need to get rid of both, do it in two steps: first paint over the date on the image, then clean the properties of the final file.

Text is removed from a phone the same way as from a computer: the online tool opens in an ordinary browser, and there is no app to install. There is one difference: a finger instead of a mouse, and that takes a little getting used to. Below are techniques for a touch screen, the built-in features of iPhone and Android, and a separate word about stickers.

Removing the date from a photo on a phone in the browser, with no app

Open the tool's page in Safari, Chrome or any other browser, tap the upload box and choose a picture from your gallery. After that everything is the same: brush, paint over, Remove. The drawing canvas responds to a finger and to a stylus.

A couple of everyday details. While you are drawing on the photo the page does not scroll, so that your finger does not drag it away instead of the brush: to scroll, swipe along the margins beside the picture. The finished preview appears below the canvas, so you need to scroll down to it a little.

How to paint over text with a finger: tips for a small screen

The main difficulty is that your finger hides the very spot you are painting. A few simple techniques help:

• turn the phone sideways if the picture is in landscape: the image fills the full width of the screen and becomes larger;
• reduce the brush with the slider to the height of the letters and move along the line in one motion, not in dabs;
• go over it twice with an overlap if you are not sure you covered the edges;
• take off any excess with the Eraser instead of starting over;
• for a small date in the corner try a stylus, or at least draw with your fingernail, which is more precise.

The brush slider has a wide range. Its value is tied to the size of the picture itself, so on a large photo from the camera the brush looks thinner on screen than on a small image from a messenger: if the stroke seems too narrow, just move the slider to the right.

How to remove text from a picture on an iPhone with built-in tools

On recent iPhones the Photos app has a Clean Up feature that removes marked objects, text included. According to Apple, it works on iPhone 16, iPhone 15 Pro and iPhone 15 Pro Max starting with iOS 18.1 and is part of Apple Intelligence, so it is not available in every region.

You will find it in the picture's editing mode: open the photo, go to Edit and choose Clean Up. Then you circle or brush over whatever is in the way with your finger. The principle is the same as in the online tool: the area is rebuilt from its surroundings, with all the same caveats about faces and patterns.

If there is no Clean Up item in the editor, it means the model or the system version does not support it, and you will not be able to turn it on with a setting.

If your iPhone has no Clean Up

On iPhone 15, iPhone 14 and older models text is removed from a picture through the browser. Open the tool in Safari, upload a picture from your library and paint over the text.

There is one subtlety with the format. By default the iPhone shoots in HEIC, while the tool accepts JPG, PNG and WebP. When you pick a picture from the library, the browser usually hands the site a JPG by itself, but if the file is taken from the Files app and has the HEIC extension, a message about an unsuitable format will appear. In that case convert the picture to JPG first, or switch the camera so that new shots are saved as JPEG straight away: Settings, Camera, Formats, the Most Compatible option.

Removing text in Samsung Gallery and in Google Photos

On Android the built-in tools depend on the manufacturer. Samsung has object removal in Gallery. According to Samsung's help, on models with Galaxy AI the order is this: open the photograph, tap Photo assist, circle the text, choose Erase, then Generate, and save the result with the Save copy button. On the Galaxy A56, A36 and A26, after Photo assist you need to choose Object eraser separately, circle what you do not want and tap Erase. On old and budget handsets the feature may be missing.

The Google Photos app has Magic Eraser. According to Google's help the route is this: open the picture, tap Edit, go to the Actions tab and choose Magic Eraser (you can also find it through search in the editor). Circle or brush over the text, tap Erase, then Done. Google lists the requirements for the phone: Android 8.0 or newer and at least 3 GB of RAM.

If your phone has neither of these, or the result did not satisfy you, the browser method works the same on any Android device.

How to remove a sticker from a photo

The answer depends on whether the picture has already been saved with the sticker. While you are still in the editor of a messenger or a social network and have not pressed Send or Save, the sticker is a separate element: you can drag it to the bin or undo the action. After saving, it is printed into the image, just like a camera date.

Look for the original first. It may still be in the phone's gallery, in the camera folder, in the chat the picture came from, or in the cloud. That is the only way to get the shot as it was before the sticker.

If there is no original, the sticker is painted over with the brush like any text, and the background is rebuilt in its place. A small emoji on a wall or on the sky will come off cleanly. A big sticker in the middle of the frame will leave a noticeable patch, because there is too much to rebuild.

What is under a sticker: why the original cannot be recovered

In a saved file there is nothing under the sticker. When saving, the messenger flattens the layers into a single picture, and the pixels the sticker covered simply never make it into the file. The neural network does not peel the sticker off or look underneath it, it draws a plausible background where it was.

So if a sticker covers a face, text, a number or anything else, there is no way to find out what was there. In place of the face you will get either a blurred patch or an invented face that has nothing to do with the real one. The same goes for marker scribbles, painted-over bars and blurred areas: you can remove the stroke itself, but you cannot bring back what is hidden under it.

Drawings and captions made in the phone's editor

Lines, arrows and captions added in the gallery's built-in editor can sometimes be undone with no neural network at all. Many galleries, including Photos on the iPhone, keep the edits separately from the original and let you bring the source picture back. On the iPhone, according to Apple's user guide, you open the edited photo, tap the button with three dots and choose Revert to Original. In other galleries look for a similar item in editing mode.

This works only if the edit was made on the same device and in the same gallery. A picture of your own that came back to you from a chat already marked up, or a screenshot with a drawn arrow that has been re-saved several times, holds no source inside. Then painting over is what is left: thin marker lines on a simple background go away nicely, bold lettering over important details leaves a mark. For more on removing objects and people by the same method, read the article on how to remove unwanted objects from a photo online.

In Photoshop text can be removed in five ways, and the choice depends on the background under the letters. For a flat background an automatic fill is enough, for a pattern and sharp lines you will need manual work. The tool names below are given as they appear in the English version of the program and in Adobe's help.

How to remove text from a picture in Photoshop: where to start

Start with a copy of the layer: duplicate the background layer or create an empty one above it and work there. That way the original image stays untouched, and a failed edit can be erased without rolling back the whole history.

Then size up the background and pick a tool. A short cheat sheet:

In practice the tools are combined: the main area is covered with a fill, and the boundaries and any repeating pattern are finished with the Clone Stamp.

Content-Aware Fill: text on a flat background

Content-Aware Fill suits cases where the text lies on a uniform surface. Select the text with any selection tool, the Lasso for example, leaving a small gap around the letters. Then open the Content-Aware Fill command in the Edit menu.

A separate workspace opens: on the left is the original image with the area the program samples from highlighted, on the right a preview of the result. If something unwanted has got into the samples (a neighbouring sign or a face, say), brush that part out of the sampling area and the preview will update. In the output settings choose New Layer: the patch will sit on top of the original, and you will be able to touch it up with a mask.

The method is fast, but on a regular pattern it produces mush, and on the boundary between two surfaces it can smear the line.

Spot Healing Brush: small text and dates

The Spot Healing Brush is handy for short pieces of text: a date in the corner, a number on a scan, a couple of words in small type. You do not need to set a sample, the program takes it by itself from around the stroke.

Choose a brush slightly wider than the height of the letters and run it over the text in a single stroke. To work on an empty layer above the original, switch on Sample All Layers in the options bar. If a blotch remains after the first stroke, go over it again with a short movement at a different angle.

On long lines and next to high-contrast boundaries the spot brush makes mistakes more often: it pulls a nearby dark or light object inwards. Then switch to the Patch tool or the Clone Stamp.

Clone Stamp: when there is a pattern or a sharp edge under the text

The Clone Stamp copies pixels from one place to another with no automation at all, which makes it irreplaceable where the pattern has to match: bricks, planks, tiles, checked fabric, the horizon line. Holding Alt (Option on a Mac), click on a clean area to set the sample, then paint over the letters.

The secret is in choosing the sample. Take it on the same line of the pattern as the spot you are painting over, and the mortar joint or the check will continue straight. Change the sample point often, otherwise visible repeats will appear. A soft brush edge hides the seams on a smooth background, a hard one is needed along sharp boundaries.

A detailed look at the settings and techniques is in a separate article on the Clone Stamp tool in Photoshop.

Patch tool: large text on a texture

The Patch tool works with a whole area: you draw around the text as with a lasso and drag the selection onto a clean area with a similar texture. The program brings the texture over from there and matches its brightness and colour to the surroundings of the original spot.

Thanks to that matching, the Patch tool is good where the lighting changes: a wall with a shadow, sand, grass, fabric with folds. Watch the preview while you drag: the lines and folds should continue. Near high-contrast boundaries the patch gives a dirty halo, so text beside the edge of an object is better removed with the Clone Stamp.

How the tool's modes and options work is covered in the article Patch tool in Photoshop: settings and use.

Generative Fill: when the background has to be drawn in

Generative Fill is available in current versions of Photoshop and is needed where a meaningful picture, not just texture, has to appear under the text: part of a shop window, a branch, the edge of a piece of furniture. Select the text, choose Generative Fill on the bar that appears and leave the prompt field empty: according to Adobe's description, in that case the selection is filled based on its surroundings. The program will offer several variations to choose from.

The feature runs through Adobe's servers and requires you to be signed in to your account and connected to the internet. The result is just as invented as with any neural network: plausible, but not what was really there. It will not bring back a face under the text either, it will draw a different one.

Recent versions also have the Remove tool: you brush over the unwanted part with it, as in the online tool, and the area is filled in automatically.

How to remove a date stamp from a photo in Photoshop

A date in the corner takes a minute to remove: zoom in on the corner, pick the Spot Healing Brush and run it over the digits. If the background under them is flat, that is enough.

If the digits have landed on parquet, a carpet or the edge of an object, first use the Clone Stamp to restore the lines that pass under the date, and cover the rest with the spot brush. Bright yellow and orange digits usually leave a coloured halo: catch it with the brush together with the digits, otherwise a barely visible warm blotch will remain in the corner.

Which to choose: Photoshop or the online tool

You need Photoshop when control matters: the pattern has to line up, the boundary has to stay straight, and the result is going to be printed in a large format. You pay for that with time and skill.

If the task is an everyday one (a date on a family photo, an inscription on a postcard, an extra line on your own screenshot), it is quicker to paint over the text in the online Remove objects from photo tool and look at the preview. If the result on a complex background does not satisfy you, that is when it makes sense to open an editor and finish it by hand.

In both cases there is one rule: work on a copy and keep the original. You may need to bring the erased text back just as suddenly as you needed to remove it.

Can I remove text from a photo online for free and without registration? Yes. No registration is needed, and processing and the preview are free: 5 edits a day from one address, a count shared by the site's neural network tools. The first one of the day starts at once, before the next ones there is a free queue of about two minutes. The preview is reduced to 900 px and carries the site's mark, the clean full-size file is paid.

Which files does the tool accept? JPG, PNG and WebP up to 25 MB, with sides from 100 to 8000 px. HEIC is not accepted, so such a picture has to be converted to JPG first. The result is saved as a JPG.

Will there be a mark where the text was? On a plain background (sky, wall, paper) the mark is usually not visible. On a face, a pattern, small print and where different surfaces meet, the patch shows, because the program rebuilds the background by guessing.

Can I find out what was under the text or the sticker? No. In a saved file there is nothing under the text or the sticker, those pixels were never recorded. The neural network draws a plausible background in that spot, it does not bring back the original content.

Why is the date still on my old photos after I turned it off in the camera? The camera setting applies only to new shots. On photographs already taken the date has become part of the image, and it can be removed only by editing: paint it over or crop it off.

What is the difference between a date on the picture and a date in the file properties? A date on the picture is digits printed into the image, and you paint over them. A date in the properties is EXIF metadata inside the file, it is not visible on the image and is removed by cleaning the metadata.

Will the photo lose quality after the text is removed? The part of the frame you did not paint over stays as it was, only the marked area changes. The file is re-saved as a JPG, and a picture with a long side over 4096 px is scaled down to that size.

What should I do if the text was removed badly? Try painting more precisely: only the letters and their shadow, with no extra area. If the background is complex, finish the spot by hand in an image editor with a clone stamp or a patch tool, or crop the frame when the text sits near the edge.