Image watermark

No, neither one. The watermark is applied inside your browser in a separate Web Worker, and the logo you select is read from your disk just like the photo itself: it is not sent anywhere and it is not stored between visits. You can confirm it by opening the Network tab in developer tools, or by disconnecting the internet after the page loads: the tool continues working normally.

Yes. The logo is drawn with its alpha channel intact, so a cropped PNG does not appear inside a white rectangle. The opacity you choose is multiplied by that transparency instead of replacing it: at 50%, an area that was already half transparent in the original file ends up at 25%, and fully transparent areas remain invisible. The logo aspect ratio is also preserved, scaling by the height you choose and calculating the width from it.

A watermark in a corner can be removed by cropping the image; a tiled one cannot. The mosaic repeats the mark across the entire canvas with a regular spacing, and if you rotate it, the whole pattern tilts with it, including the corners: the grid extends beyond the edge on purpose so the rotation does not leave gaps. This is useful when you share an original that you do not want to circulate without attribution; a corner mark is more discreet and interferes less with the photo.

Yes, and it is not an approximation. The preview and the processing share the same function for placing the stamp: positions are calculated as fractions of the canvas, so the only difference between what you see and what you download is the number of pixels. Position, relative size and angle are identical.

Because the work happens in a Web Worker, which is a separate thread and does not share the web fonts loaded by the page. If we allowed a specific typeface, the browser would substitute another without warning and the downloaded file would not match the preview. The three options —sans-serif, serif and monospace— are generic families that the system always resolves.

There is no deliberate loss, but there is re-encoding: the result passes through the browser encoder again. With lossless PNG and WebP there is no visible degradation; with JPG it is re-encoded at a high fixed quality, so the difference is minimal. The output format and size remain the same as the input.

Yes. Upload the whole batch, set the watermark once and click process: the same text or logo is applied to every image in sequence, with progress visible while it runs. Because position and size are stored as fractions rather than pixels, the stamp remains proportionate whether the photo is 800 px or 6000 px, without needing to adjust anything between them.

No. All processing happens on your device, so we do not use our infrastructure and there is no cost or limit on how many images you can watermark.