Images
How to Add a Watermark to Photos and PDFs (Protect Your Work Without Ruining It)
July 11, 2026 · 8 min read
A watermark's job is not to make your image ugly. It is to make reuse annoying enough that someone credits you, licenses the file, or moves on. The whole game is finding the setting that is visible enough to discourage lifting but light enough that people still enjoy the photo.
For most photos the answer is a semi-transparent mark at roughly 15 to 30 percent opacity, sized to about a tenth of the image, placed where it cannot be cleanly cropped off. For documents you do not cover the pixels at all: you stamp text like DRAFT or your name diagonally across each page. The mechanics differ between a JPG and a PDF, and this guide covers both.
One honest note up front: no watermark is theft-proof. A determined person with a content-aware fill tool can wipe a small corner logo in seconds. What a watermark reliably does is deter casual copying, assert ownership, and advertise you when the image travels. Set your expectations there and every other choice below gets easier.
What a watermark does, and what it can't do
A visible watermark works by raising friction. Someone who wanted to grab your photo for a quick blog post or marketplace listing now has to either credit you, crop and clone it out, or find a cleaner source. For a lot of casual reuse, that friction is enough. It also does two quieter jobs: it stamps a claim of authorship on the file, and it turns every re-share into a small billboard for your name or domain.
What it will not do is stop a motivated thief. Corner marks get cropped. Low-opacity logos get painted out with content-aware fill in a modern editor. And a watermark does nothing against reverse image search, screenshots, or someone simply retyping your text. Treat it as a deterrent and a signature, not a lock.
- Good for: deterring casual reuse, crediting yourself when images spread, marking proofs and previews before payment.
- Not good for: stopping determined removal, protecting against reverse image search, or replacing a proper license or copyright registration.
- Belt and suspenders: keep your untouched originals and your camera or export metadata. If ownership is ever disputed, the full-resolution original with intact EXIF is far stronger evidence than a watermark.
Text or logo? Choosing the mark itself
A text watermark, your name, @handle, or domain, is the workhorse. It is legible at any size, it scales cleanly, it doubles as free advertising, and even if someone crops the image they may still remember the words. It is the right default for photographers, sellers, and anyone whose goal is attribution.
A logo watermark reads as more polished and builds brand recognition over time, which is why studios and agencies lean on it. The catch: a logo carries no readable information once it is small and faint in a corner, so if it gets cloned out there is nothing left to trace you by. When you add a logo watermark to an image, keep a version with your name or URL nearby, or bake the two together, so the mark still says who made it.
The strongest compromise is a combined mark: a small logo next to your handle or site, exported once as a transparent PNG so you can drop it onto everything consistently. Consistency matters more than cleverness, one recognizable mark used everywhere beats a different treatment on every post.
- Text mark: maximum legibility and attribution, zero design work, best for individuals and sellers.
- Logo mark: brand recognition and a professional look, best when the brand is already known.
- Combined mark: logo plus handle in one transparent PNG, the most resilient option and the easiest to reuse.
The three settings that decide everything: opacity, size, color
Almost every ruined watermark fails on one of three dials. Get these right and even a plain text mark looks intentional.
Opacity is the big one. A single mark usually lands well between 15 and 30 percent, visible on a second look but not fighting the subject. If you are tiling the mark across the whole frame, drop each instance to roughly 8 to 15 percent, because there is far more of it on screen and the coverage does the deterring, not the darkness. Above about 40 percent a single mark starts to feel like graffiti; below about 10 percent it disappears against busy areas.
Size should be proportional, not fixed. A corner mark around 10 to 20 percent of the image width reads clearly without dominating. Setting a pixel size instead means the same watermark is huge on a thumbnail and invisible on a 4000-pixel export. Scale it to the image.
Color is really about contrast. Plain white or plain black will vanish wherever the photo happens to match it. The fix is to give the text a thin outline or a soft drop shadow so it stays legible over both bright skies and dark shadows. White text with a subtle dark shadow is the most reliable single choice across mixed images.
- Single mark opacity: 15 to 30 percent. Tiled mark opacity: 8 to 15 percent per instance.
- Size: 10 to 20 percent of image width for a corner mark, always proportional, never a fixed pixel value.
- Legibility: add a thin outline or drop shadow so the mark survives both light and dark regions of the photo.
Placement: corner, center, or tiled
Where you put the mark is a direct trade between protection and how much you intrude on the image. The right answer depends on whether the picture is a finished piece you want people to admire or a proof you do not want reused at all.
A single corner mark protects the look but is the easiest to defeat: a crop or a quick clone removes it. A centered or diagonal mark is much harder to erase cleanly but sits on top of your subject. Tiling, repeating the mark across the whole frame, is the most theft-resistant because there is nothing to crop to, and it is the standard for client proofs and any image where a leak needs to be traceable. Match the placement to how sensitive the image is.
| Placement | Deters theft | Gets in the way | Best for |
|---|---|---|---|
| Single corner | Low to medium (easy to crop) | Minimal | Portfolios and social posts where the look matters most |
| Center, single mark | Medium to high | High | Proofs and previews you do not want reused as-is |
| Diagonal across | High | Medium to high | Stock-style previews and sample images |
| Tiled / repeated | Highest (nothing to crop to) | High if too dark | Client proofs, sensitive images, and leak tracing |
How to watermark a photo, step by step
The workflow is the same whether you use a full editor or a quick browser tool. If you would rather not install anything, you can add a watermark to a photo directly in the browser and the file never leaves your machine.
The steps below assume a text or transparent-PNG mark and a single JPG or PNG. For a whole folder, look for a batch mode so you apply identical settings in one pass rather than eyeballing each file.
- Start from a copy, never your only original. Watermarking re-encodes the file, and you want the clean master preserved.
- Load the photo and add your text or drop in your transparent logo PNG.
- Set opacity to about 20 percent to start, then nudge until it is readable but not distracting.
- Scale the mark to roughly 10 to 20 percent of the width and position it, corner for finished work, tiled for proofs.
- Add a thin shadow or outline if the mark disappears over any part of the image.
- Export at high quality (JPG quality around 85 to 90) to avoid stacking compression artifacts, then check it at full size before publishing.
Watermarking a PDF is different
A PDF is not a single picture, so you do not overlay pixels the way you do on a photo. Instead you stamp text, DRAFT, CONFIDENTIAL, your name, a date, across every page at once, usually as faint diagonal type behind or over the content. Because it is real text, it stays razor-sharp at any zoom and adds almost nothing to the file size. You can watermark a PDF this way in one operation instead of touching each page.
This is the format for contracts, proposals, proofs, and anything circulating before it is final. Standard practice is light gray diagonal text at low opacity so the document stays readable underneath. If you need the mark to survive printing and re-scanning, make it a touch darker, since a very faint stamp can wash out on a photocopy.
The same caveat applies as with photos: a PDF stamp can be removed by someone with the right editor, and a low-opacity mark over a scanned image page behaves more like an image overlay than crisp text. Watermarking signals status and ownership; it does not lock the document. For genuine access control, pair it with a password or restricted permissions rather than relying on the stamp alone.
- Photo watermark: pixels overlaid on a raster image, tuned per image, best exported as a copy.
- PDF watermark: text stamped across every page, stays vector-sharp, ideal for DRAFT and CONFIDENTIAL marking.
- For real protection on a PDF, combine the watermark with a password or permission restrictions, not the stamp by itself.
Common mistakes, and the fixes
Most watermarking regret comes from a handful of avoidable errors. Run through this list before you publish a batch you cannot easily redo.
- Watermarking your only copy. Always work from a duplicate so the clean original survives.
- Cranking opacity too high. If the mark is the first thing people see, it is too strong, drop back toward 20 percent.
- Using pure white or pure black with no outline, so the mark vanishes over matching areas of the photo.
- Placing a small logo in a corner on a valuable image, it is the single easiest thing to crop or clone away.
- Re-saving a JPG several times at low quality, which stacks visible artifacts; export once at high quality from the master.
- Forgetting to compress for the web afterward. A finished, watermarked image is often larger than it needs to be, shrink it before uploading so pages stay fast without visibly degrading the mark.
Frequently Asked Questions
It stops casual reuse, not determined theft. A visible mark makes lifting your image annoying enough that many people credit you, license it, or move on, and it advertises your name wherever the image travels. But a corner logo can be cropped, and a faint mark can be painted out with content-aware fill in seconds. Treat it as a deterrent and a signature, and keep your full-resolution originals as your real proof of ownership.
For a single mark, somewhere between 15 and 30 percent usually works: visible on a second look but not fighting the subject. If you tile the mark across the whole image, lower each instance to about 8 to 15 percent, because the coverage does the work, not the darkness. Above roughly 40 percent a single mark starts to look like graffiti; below about 10 percent it disappears over busy areas of the photo.
Corners protect the look of the image but are the easiest to crop or clone out, so they suit finished portfolio and social posts. A centered or diagonal mark is much harder to remove cleanly but sits on your subject, which is fine for proofs and previews. Tiling the mark across the whole frame is the most theft-resistant because there is nothing left to crop to, which is why it is standard for client proofs and sensitive images.
Sometimes easily. A small corner mark can be cropped away or erased with a content-aware fill tool in a modern editor, often in one step. Centered, diagonal, and tiled marks are much harder to remove without visibly damaging the image. Nothing is bulletproof, though, so keep your untouched originals and intact metadata, which are far stronger evidence of ownership than the watermark itself if a dispute ever arises.
On a photo you overlay actual pixels on a raster image, tuned per picture, and export a copy. On a PDF you stamp real text, like DRAFT or CONFIDENTIAL, across every page at once. That PDF stamp stays sharp at any zoom, barely changes the file size, and is applied in a single pass. For a document you also usually want the mark faint and diagonal so the content stays readable underneath.
Watermarking re-encodes the file, so work from a master copy and export once at high quality (JPG around 85 to 90) to avoid stacking compression artifacts. Quality-wise the mark itself does not degrade the photo if you apply it cleanly. For speed, the finished file is often larger than it needs to be, so compress it before uploading so your pages stay fast while the watermark still reads clearly.
Export your logo as a transparent PNG so only the mark shows, then drop it onto the image and scale it to about 10 to 20 percent of the width. Because a logo carries no readable name, pair it with your handle or domain, either beside it or baked into the same PNG, so the mark still identifies you if part of the image is cropped. Reuse that one file everywhere for a consistent, recognizable brand.
Related Tools
Overlay text or a transparent logo on any JPG or PNG with opacity and placement controls, right in your browser.
add a watermark to a PDFStamp DRAFT, CONFIDENTIAL, or your own text diagonally across every page of a document in one pass.
compress the watermarked imageShrink the finished file for fast web upload without visibly degrading the mark you just added.