💧 Image Watermark
Batch watermark product photos: text or logo, nine-grid positioning or full-page tiling, adjustable opacity and angle, ZIP batch export; optional hidden tracking marker and metadata, verifiable after export; processed locally in your browser.
No file imported. Showing the fictional sample.
💧 Watermark settings
Watermark typeOutput settings
👁️ Live preview
No file imported yet. Showing the sample.
Thumbnails will appear here after you import images
🔍 Tracking Mark & Verification
Experimental DCT blind watermark: invisible to the eye, written into the image pixels at export; still helps verify the number if the metadata is deleted.
Writes the purpose watermark, tracking number, export settings, and time into the PNG / JPEG file header; recipients can read and delete it.
When on, the number is “20261106T053000Z-random-number”, using your computer’s clock; when off, only the 16-hex number is used.
The full number is generated at export.
Generate a different number for each recipient and keep your own recipient–number list; the number contains no name, account, or device info.
⚠️ A hidden mark is not a digital signature: cropping, scaling, screenshots, or heavy compression may break it, and numbers can be forged — it cannot by itself prove who shared a file.
Verify tracking mark
Choose a PNG or JPEG exported by this tool to read its metadata and hidden pixel mark locally; metadata alone does not mean the hidden mark is still valid.
No file selected for verification yet.
No file selected for verification yet.
Tracking records
No records exported yet this session. Records stay only in this page's memory and are cleared when the page closes; they are never uploaded or saved to local storage.
How to use Image Watermark
Follow the steps to get it done
Pick or drag in images (PNG/JPEG/WebP, up to 30)
Open the page and start — no registration or sign-in needed.
Choose a text or Logo watermark, adjust its position, opacity, and angle, and check the live preview
Settings can be changed at any time; all computation happens in your browser.
Turn on the tracking mark if needed, then save single images as PNG / JPEG or export the whole batch as one ZIP
Copy or download the result when you are done. Nothing is kept on this site.
Page Actions
Click to open the system file picker. You can select several PNG/JPEG/WebP images (hold Ctrl to pick several, or Ctrl+A to select all, up to 30).
Drag image files from your computer straight onto the upload area and drop them to add them in bulk.
The “+ Add” button in the toolbar adds more images to the existing list (up to 30). You can also drag new images onto the uploaded list to append them.
The view switch in the toolbar: grid mode shows image thumbnails directly, list mode shows file names with image sizes. List mode is the default. The grid is a single horizontal row; when files exceed the width, use the ‹ › arrows on either side of the display area to scroll (dragging a card reorders it, it does not scroll).
Each image entry has a red ✕ at its top right corner. Click it to delete that image directly, with no confirmation. The list and the preview update together. When the last image is deleted, the page returns to the fictional sample automatically.
Two icon buttons on each image entry (they show only the ↑ ↓ arrows, no text; the title and aria-label are kept). One click moves that image one step up or down in the preview and batch export order. They work alongside drag sorting; move-up is disabled on the first image, move-down on the last.
Press and hold an image in the uploaded list and drag it, then release to change its order in the preview and batch export. After dragging, the preview stays on the image you just moved.
Click an image in the uploaded list and the live preview on the right switches to it immediately.
Below the live preview is a strip of thumbnails for all images (with number badges). Click any one to switch the large preview at once; the current one has a blue border.
Adjusts the size of the thumbnails in the strip below the preview: drag the slider, click − / +, or type a value from 50–150 directly in the number box (out-of-range values are clamped automatically and a note appears in the status area). It only affects how the thumbnails are displayed, not the export.
Above the large preview it shows which image this is and its file name, updating as you switch images.
A summary line stays below the large preview: which image this is out of how many, the file name, the original size, the preview size, the watermark type, and the opacity.
Two pill buttons switch the watermark type: a text watermark uses custom wording, a Logo watermark uses an uploaded transparent image.
Multiple lines of text, with variables: %date% for the current date, %time% for the current time, and %filename% for the original file name (without extension), replaced automatically when rendered.
14–80 px (measured at the preview size), controlling the size of one line of watermark text. At export it is scaled up proportionally to the original image. A number box sits next to it for typing a value directly, kept in sync with the slider; out-of-range values are clamped automatically and a note appears in the status area.
Normal or bold, controlling the stroke weight of the watermark text.
Sans-serif, serif, or monospace system fonts, switching the watermark text's font family.
Click the color swatch to pick the watermark text color. The default is white; use a dark color on light photos.
10–100%, shared by text and Logo watermarks. The smaller the value, the fainter the watermark; around 60% works well for photos. A number box sits next to it for typing a value directly, kept in sync with the slider; out-of-range values are clamped automatically and a note appears in the status area.
In Logo watermark mode, click this button to upload a PNG or SVG image (a transparent background works best). A small thumbnail appears after upload.
Removes the uploaded Logo image, and the thumbnail disappears. It does not affect the images already imported for processing.
5–50%, the Logo width as a percentage of the original image width. The default is 15%. A number box sits next to it for typing a value directly, kept in sync with the slider; out-of-range values are clamped automatically and a note appears in the status area.
Two pill buttons switch the layout: single-point puts the watermark in one position, full-page tile covers the whole image with watermarks.
In single-point mode, pick where the watermark sits on the image: top-left, top-center, top-right, middle-left, center, middle-right, bottom-left, bottom-center, or bottom-right. The default is bottom-right.
In single-point mode, fine-tunes the watermark's horizontal position, -20%–20% (relative to the image width). Positive values move it right, and it sticks to the edge if it goes past it. A number box sits next to it for typing a value directly, kept in sync with the slider; out-of-range values are clamped automatically and a note appears in the status area.
In single-point mode, fine-tunes the watermark's vertical position, -20%–20% (relative to the image height). Positive values move it down, and it sticks to the edge if it goes past it. A number box sits next to it for typing a value directly, kept in sync with the slider; out-of-range values are clamped automatically and a note appears in the status area.
In full-page tile mode, the rotation angle of the watermark, -60°–60°, default -24°. It has no effect in single-point mode. A number box sits next to it for typing a value directly, kept in sync with the slider; out-of-range values are clamped automatically and a note appears in the status area.
In full-page tile mode, the horizontal gap between neighbouring watermarks, 0.5–5 times the font size (in Logo mode, a multiple of the Logo width). A number box sits next to it for typing a value directly, kept in sync with the slider; out-of-range values are clamped automatically and a note appears in the status area.
In full-page tile mode, the vertical gap between neighbouring watermarks, 0.5–5 times the font size (in Logo mode, a multiple of the Logo width). A number box sits next to it for typing a value directly, kept in sync with the slider; out-of-range values are clamped automatically and a note appears in the status area.
Keep the original size, or cap the output's longest side at 1600/2400/3200 pixels. Useful for shrinking very large originals.
The image format for batch ZIP export: JPEG or PNG. The single-image save buttons are not affected by this setting.
The file name rule for batch export: the original file name with a _watermarked suffix, or sequential names like image_01, image_02, and so on. When the tracking switch is on, both rules also append -tracking-number before the extension.
60–98%, only effective for JPEG export; PNG is lossless and unaffected. A number box sits next to it for typing a value directly, kept in sync with the slider; out-of-range values are clamped automatically and a note appears in the status area.
When several images are imported, switch the image shown in the live preview.
Exports the current preview image (with watermark) as a lossless PNG and downloads it. When the tracking switch is on, the hidden mark is embedded and the metadata is written at the same time, and the file name carries the tracking number.
Exports the current preview image (with watermark) as a JPEG at the selected quality and downloads it. When the tracking switch is on, the hidden mark is embedded and the metadata is written at the same time, and the file name carries the tracking number. After JPEG compression, the mark is decoded and verified; if verification fails, the save stops outright.
The “📦 Bulk export ZIP (all images)” button watermarks every imported image one by one and packs them into a single ZIP to download. A progress bar appears during export, and you can click “Cancel” at any time to stop the remaining work. When the tracking switch is on, the whole batch shares one tracking number, embedded in each image with metadata written and the number in the file name, and the ZIP adds one tracking record.
Click this button during a batch export to stop the images that have not started yet. Images already processed are not affected.
The first item in the tracking panel, on by default: at export, the tracking number is written into the image pixels (a DCT blind watermark, invisible to the eye). Turn it off and exports carry no mark. If an image is too small (about smaller than 176×176 pixels) or the picture is badly over- or underexposed, the mark may not embed, and the export errors out directly instead of silently producing an unmarked image.
On by default: at export, the usage watermark, tracking number, export parameters, and time are written into the PNG / JPEG file header. The recipient can read and delete them; if deleted, the hidden pixel mark can still help with cross-checking.
On by default: the number takes the form “20261106T053000Z-random-number”, with the time taken from this machine's clock (UTC). Turn it off and only the 16-digit random number is used.
16 hexadecimal digits, editable by hand. The format is checked at export, and an invalid value errors out and stops the export. A new one is generated automatically each time the page is opened, and it is not saved in the settings cache.
Click “Generate a new number” once to generate a new 16-digit number. Generate a new one before exporting for each different recipient, so you can tell them apart later by number.
A line of text stays below the number input box. Before exporting it reads “The full number is generated at export.”, and after exporting it shows the full number actually used for that export.
Click to pick one PNG or JPEG (a single file no larger than 160 MiB). The metadata and the hidden pixel mark are read locally and separately, and the results go into the verification area below. The file is not uploaded.
The verification area always shows three conclusions: the metadata number and usage watermark, the hidden pixel mark number and time, and whether the two agree. If only metadata is found, it notes “Metadata alone does not mean the hidden mark is still valid”.
The “💾 Save tracking records” button downloads this page's export records as watermark-trace-records.json (each record holds the file name, tracking number, SHA-256, and whether metadata was written). It is disabled when there are no records. Records live only in the current page's memory, and closing the page clears them.
The “Clear” button in the toolbar: click once and it turns into “Confirm clear?”, click again to remove all imported images and return to the fictional sample state to start over. If you do not confirm within 5 seconds, the guard drops automatically.
How It Works & Highlights
How it works, and where it fits
This tool draws watermarks locally in the browser with Canvas: it first reads the image and corrects the orientation automatically from the EXIF data (phone photos do not come out upside down), then draws the text or Logo into the pixels at the watermark type, position, opacity, and angle you set. The preview redraws a scaled-down image live, and at export the image is redrawn on an offscreen Canvas at the original image's true resolution, with every font size, spacing, and Logo size scaled proportionally by 'export width ÷ preview width', so the watermark's visual proportions match the preview exactly. Batch export processes up to 3 images at a time, releasing memory image by image, and finally packs the results locally with JSZip for download. Below the preview area is a thumbnail strip; click one to switch the large image. The upload list supports drag sorting, and the export order follows it. A watermark is drawn into the pixels, not a separate layer, so it cannot be removed on its own after export. All processing happens locally, with no upload. Tracking uses a DCT blind watermark from a reference implementation (offline document watermark v0.5.0, MIT): the number (a 16-digit random value, with an optional UTC time prefix) is encoded into a data packet with a magic number and CRC check, written bit by bit into the luminance of 8×8 pixel blocks through two mid-frequency DCT projections, each bit repeated across several blocks and recovered by voting. At the same time, the number can be written into the PNG iTXt block or the JPEG COM segment metadata, and carried into the file name and the tracking record. The preview carries no mark; the mark is added only at export. After JPEG export, the compressed result is decoded and verified, and the export stops outright if the mark was damaged by compression. During verification, the metadata and the pixel mark are read locally and separately, and the two numbers must agree.
Highlights
- ✓ Text and Logo watermarks, 3×3 grid positioning or full-page tiling, with a live preview as you change settings
- ✓ Phone photos are corrected automatically from EXIF data, batch ZIP export in one go, with cancel support
- ✓ Export redraws at the original resolution, so the watermark proportions match the preview exactly
- ✓ Hidden tracking mark plus metadata: exports are traceable, with local number verification
Tips
- 💡 Before sending a product photo to a customer, use a single-point Logo watermark at the bottom right with about 60% opacity: it deters theft without stealing focus from the subject
- 💡 To guard against screenshots and cropping, use full-page tiling with a -24° tilt. Cropping out one part of the image still will not remove it cleanly
- 💡 Use the %filename% variable in the watermark text, and each image in a batch export automatically carries its own file name in the watermark
- 💡 Before sending product photos to different customers, click “Generate a new number” and then export, so the number tells you later which export an image came from. Remember to click “Save tracking records” to keep your own records
FAQ
Privacy, format limits, and usage tips