Add text or logo watermarks to whole batches of photos — fonts, opacity, blend modes, tiled or custom placement — and download as a ZIP.
Content last reviewed
Drop images to watermark
JPG, PNG, WEBP (Batch supported)
Drop in one image or a whole batch to watermark together.
Type text (or upload a logo), then set font, size, colour, opacity, blend mode, and position.
Judge the live preview, then download the watermarked set — batches arrive as a ZIP.
The standard portfolio setup: subtle, consistent, applied to every file at once.
Input
18 photos · Text "© 2026 Mara Lindqvist" · bottom-right · 40% opacity · size 4% of widthOutput
18 files with the mark scaled per image, downloaded together as watermarked-images.zipRepetition across the surface makes clean removal impractical — the proof convention.
Input
Draft photo · Text "PROOF" · Tiled · 30% opacity · rotatedOutput
The image covered in a repeating diagonal PROOF pattern, unusable as a final but clearly reviewableA watermark is a small ownership claim that travels with the image: your name, brand, or logo placed over a photo so that when it is shared, screenshotted, or reposted, the attribution rides along. This tool applies one to a whole batch at once — text or an uploaded logo — with the styling control to make it look deliberate rather than stamped on.
Text marks draw from ten embedded font choices with size set as a percentage of each image's width (so the mark scales sensibly across a mixed batch), plus colour, bold, italic, and rotation for the classic diagonal. Logo marks scale the same relative way. Placement covers the five standard spots — four corners and centre — plus two that matter for protection: tiled, which repeats the mark across the whole surface, and custom, where you drag the mark to any position on the preview. Opacity (default 50%) and five blend modes control how the mark sits over both light and dark regions, and everything re-renders live so you judge on the actual photo.
Be realistic about what a watermark does: it deters casual reuse and keeps attribution attached — it does not stop a determined remover. That is why the working convention is corner marks at 30–50% opacity for branding, centre or tiled placement for proofs and previews, and a clean un-watermarked master kept archived. Your photos are processed on your device and never uploaded; batch results download together as a ZIP.
Drop your images onto the upload zone — the whole batch shares one watermark configuration.
Choose the watermark type: Text (type your mark — a © with the year reads well) or Image, uploading a logo file.
Style it: for text, pick one of the ten fonts, set the size (as % of image width), colour, bold/italic, and rotation; for a logo, set its scale. Then choose opacity (start around 40–50%) and a blend mode.
Pick the position: a corner, centre, Tiled to repeat across the image, or Custom and drag the mark on the preview.
Check the live preview — switch between batch images to confirm the mark works on light and dark photos alike.
Click download: a single image saves directly, a batch arrives as watermarked-images.zip. Keep the clean originals archived.
Type a mark or upload a logo image — both scale relative to each photo's width.
One configuration applies across all uploaded images, with per-image preview and a ZIP download.
Font choice, size as a percentage of width, colour, bold, italic, and rotation for diagonal marks.
Four corners, centre, tiled repetition across the surface, and free drag-to-place custom positioning.
An opacity slider plus Normal, Multiply, Screen, Overlay, and Soft Light blending for legibility over any background.
The watermark re-renders on the actual photo as you adjust, so judgment happens before download.
Brand photography and illustrations before they circulate on social platforms.
Mark preview batches centre or tiled so drafts cannot be used as finals before payment.
Apply a consistent logo mark across a whole shoot in one pass.
Stamp marketplace or property images so competitors cannot lift them unattributed.
Each photo is drawn onto a canvas at full resolution, then the watermark is composited over it with your opacity and blend mode applied. Text marks are rendered at a font size computed from the image's width (which is why one size setting works across a mixed batch) after the chosen font is loaded and ready; logo marks are scaled the same relative way. Position presets compute the placement with a sensible margin; tiled mode repeats the mark in a grid across the whole canvas; custom mode uses the coordinates you dragged to on the preview.
The preview and the final render share the same drawing code, so what you approve is what exports. Finished images are encoded as high-quality JPEGs named after the original with a _watermarked suffix, and batches are zipped in the browser.
JPG, PNG, WebP and other browser-decodable images for the photos; any image file for the logo mark.
Watermarked images export as high-quality JPEG with a _watermarked filename suffix; batches download as watermarked-images.zip.
Text with font, size %, colour, bold, italic, and rotation — or an uploaded logo with a scale control.
Top-left, top-right, centre, bottom-left, bottom-right, tiled repeat, or custom drag placement.
0–100% transparency of the mark, defaulting to 50%.
Normal, Multiply, Screen, Overlay, or Soft Light compositing.
A white mark over a bright sky, or a dark one over shadow, disappears. Switch the blend mode — Overlay and Soft Light adapt to the underlying tones — or nudge the colour and opacity.
Output is JPEG, which has no alpha channel, so transparent PNG sources gain a solid background. Watermark flattened copies, and keep transparent masters as PNG.
Text fonts load from Google Fonts; if that request is blocked (offline, network filter), drawing falls back to a default font. Retry online or pick another font.
Output is JPEG-only, so transparency in source images is flattened and lossless output is not available. One watermark configuration applies to the whole batch — no per-image variation within a run. Text fonts are fetched from Google Fonts, which requires a network connection for the font files (the photos themselves never upload). And the perennial honest note: a watermark deters and attributes, but it is not DRM — determined removal, especially with modern tooling, remains possible, which is why originals should be archived clean.
Your photos are composited on a canvas in your browser and never uploaded — the batch ZIP is also built locally. The one network activity involved is the font files loading from Google Fonts when you use a text watermark; your image data does not travel with that request.
Keep branding marks at 30–50% opacity — visible when looked at, ignorable when viewing the photo.
Use centre or tiled placement for proofs; corner marks are croppable by design.
Preview against the brightest and darkest photos in the batch before downloading.
Archive clean, un-watermarked masters — the watermarked file is a distribution copy.
Prefer Overlay or Soft Light blends when a batch mixes light and dark images.
Decide the mark's job first. Branding wants small, consistent, corner-placed, and subtle — the viewer should register the image first and the mark second. Protection (proofs, previews, drafts) wants centre or tiled placement at higher opacity, where removal costs more than paying for the clean file. The same tool serves both; the settings should not be the same.
Standardize once: pick your font, size percentage, opacity, and corner, and reuse them across every batch. A consistent mark reads as identity; a different mark per upload reads as noise.
And maintain the two-copy discipline — clean master archived, watermarked copy distributed. The watermark is for the copy that leaves your hands; printing, licensing, and future edits all come from the master.
Rendering is canvas-based at source resolution: the photo draws first, then globalAlpha (opacity) and globalCompositeOperation (blend mode) are set and the mark is drawn — text via fillText at a font size computed as sizePercent × image width after awaiting document.fonts loading, logos via drawImage at the scale percentage. Tiled mode steps the mark across the canvas in a grid; custom mode maps preview-drag coordinates to image space. Output encodes as image/jpeg at quality 0.92 with a _watermarked suffix; batches are assembled with JSZip client-side. Fonts load from the Google Fonts CDN; image data stays local.
| Placement | Removal resistance | Typical use |
|---|---|---|
| Corner (any of four) | Low — croppable | Everyday branding |
| Centre | Medium — over the subject | Proofs, previews |
| Tiled | High — covers everything | Drafts, stock-style protection |
| Custom drag | Depends on placement | Awkward compositions |
Photographers and illustrators branding portfolio posts, studios issuing client proofs, shops stamping product and listing images, and anyone who publishes visual work and wants attribution to survive a repost. It handles the batch case well; for a single quick mark it is equally at home.
Match placement to purpose: bottom-right (or any corner) at modest opacity for branding — professional but croppable; centre or tiled for proofs and previews, where removal has to be harder than paying for the clean version. The tiled mode exists exactly for that second case.
30–50% for branding marks — clearly legible when you look, unobtrusive when you are viewing the photo. Push higher only for proofs where deterrence matters more than aesthetics; fully opaque marks read as heavy-handed.
Yes — switch the watermark type to Image and upload the logo. It scales as a percentage of each photo's width, and all the same opacity, blend, and position controls apply.
Copyright exists automatically when you create the image — the watermark adds visible attribution, deterrence, and evidence. Notably, removing someone's watermark is itself unlawful in many jurisdictions under copyright-management-information rules. It strengthens your position; it is not what creates the right.
No — compositing happens on a canvas in your browser and the ZIP is built locally. The only network fetch is the Google Fonts files when a text mark uses one; your images never travel.
Output is JPEG, which cannot store transparency, so alpha regions are flattened. Keep transparent masters as PNG and watermark flattened distribution copies.