Making Your Own Text Watermark for Free

A watermark is just a transparent PNG placed on top of a photo. You do not need a subscription, a desktop app, or a plugin — you need one well-made overlay file that you reuse forever. The craft is entirely in getting that one file right, and in placing it somewhere a thief cannot simply crop away.

Decide what the watermark is for first

People conflate two very different goals, and they call for opposite designs.

Attribution means someone reposting your work should be able to find you. It wants a small, tasteful mark — your handle or domain in a corner. It accepts being removable, because you are relying on goodwill rather than force.

Deterrence means the image should be unusable commercially without your permission. It wants a mark across the centre of the frame, large and overlapping the subject, so removing it destroys the image. Photographers use this on client proofs.

Pick one deliberately. A small corner mark provides essentially no deterrence — it takes seconds to crop. A big central mark ruins the viewing experience for the audience you actually want. Trying to do both at once produces a watermark that is simultaneously annoying and ineffective.

Building the overlay file

Open our Text to Image Converter and work through these settings in order. The whole thing takes about two minutes.

Step 1: Write the shortest useful text

Your domain (yoursite.com) or handle (@yourname). One line. Resist adding "© 2026 All Rights Reserved" — it doubles the width, halves the legible size, and carries no extra legal weight. Copyright exists whether you print the word or not. A domain is more useful anyway, because it tells a viewer where to go.

Step 2: Turn on transparent background

Non-negotiable, and it must be enabled before you generate rather than removed afterwards. Otherwise your letters get anti-aliased against white and you will see a pale halo around every character once the mark sits on a dark photo. We cover exactly why in the transparency guide.

Step 3: Choose white, and go bold

White works on more photos than black, because most photographs have more midtone and shadow area than blown-out highlight. Use a bold or semi-bold weight: thin strokes lose too much of their visual weight once you drop the opacity, and they disappear entirely over textured areas.

Step 4: Generate far larger than you need

Set a font size around 120px even for a small corner mark. You will scale the overlay down when you apply it, and downscaling keeps edges crisp. Scaling a small watermark up produces soft, blocky letters that look amateurish and, ironically, cheapen the photo you are trying to protect. One large master file serves every output size you will ever need.

Step 5: Add generous padding

Padding gives the anti-aliased edges room to breathe and makes the file easier to position later without accidentally clipping a descender. It costs almost nothing in file size on a transparent PNG, since empty pixels compress to nearly nothing.

Download the PNG and store it somewhere permanent alongside your other brand assets. This single file is now your watermark for every photo you publish.

Applying it without ruining the photo

Opacity is the whole game. A watermark at full opacity competes with your subject and reads as clutter. Somewhere between 25% and 45% is usually right: present enough to be legible if someone looks for it, quiet enough to ignore while viewing the image. Test on your darkest and lightest photos, not a mid-grey one.

Scale it to roughly 15–20% of the image width for a corner mark. Smaller becomes illegible once a platform compresses the image; larger starts dominating.

Never place it flush against the edge. Leave a margin of about 3% of the image width. Edge-flush marks look accidental, and platforms sometimes crop a few pixels for thumbnails — taking your watermark with them.

Placement, and the cropping problem

The bottom-right corner is conventional, which is exactly why it is the first place anyone crops. If attribution genuinely matters to you, think about this properly.

Over the subject, not the background. A mark that overlaps the main subject cannot be cropped without cutting into what makes the photo worth stealing.
Off-centre and slightly rotated. Awkward positions are harder to clone out cleanly than a mark sitting on flat sky.
Tiled across the frame for proofs and previews. Repeating the mark at low opacity in a diagonal grid makes removal genuinely impractical. Reserve this for images you do not intend people to enjoy.
Avoid flat areas. A watermark on plain sky or a white wall is trivial to paint over. Busy texture is much harder to reconstruct.

Be realistic about all of this: a determined person with modern editing tools can remove almost any watermark. The purpose is to make casual, careless reuse inconvenient and to ensure honest people can credit you — not to build an unbreakable lock.

Batch-watermarking many photos

Once you have the overlay, applying it to a hundred photos should not mean a hundred manual steps. Any tool that supports layers or actions can do it: place the overlay once, record the action, then run it over a folder. Our Merge Images tool handles the composite step in the browser if you would rather not install anything.

Two rules for batch work. First, always keep your originals — write watermarked copies to a separate folder, because a watermark cannot be removed from a flattened export. Second, check a sample from each lighting conditionbefore running the whole set. A white mark tuned on bright outdoor shots will vanish on your night photography, and you will not notice until all two hundred are done.

Common mistakes

A watermark bigger than the message. If the first thing people see is your logo, the photo has become an advertisement for itself.
Illegible after compression. Social platforms re-compress everything. Thin, small, low-contrast marks turn to mush. Check how yours survives an actual upload.
Watermarking work that is not yours. Adding your mark to stock images or someone else's photo misrepresents authorship, regardless of intent.
Inconsistent marks. Different fonts, sizes and positions across your feed make the mark read as noise rather than a signature. One file, used the same way, builds recognition.
Relying on metadata alone. EXIF copyright fields are worth filling in, but nearly every platform strips them on upload. A visible mark survives; metadata usually does not.

Make yours now

Head to the Text to Image Converter, type your domain, tick transparent background, set the size high, and download. Everything happens in your browser, so neither your text nor your photos are ever uploaded anywhere.