Watermark Image

Put a text or logo mark on an image, at a size, colour, angle and opacity you set.

Runs on your deviceImagesWorkingNo account needed

What this tool does

Watermark Image draws the mark straight onto the picture's pixels. The photo is decoded into a canvas, the text or logo is composited over it at the opacity you give, and the canvas is encoded as a new file with -watermarked added to the name. Text is measured and drawn by the browser's own type engine in one of three families — sans-serif, serif or monospace — and sized as a percentage of image width, so one setting looks the same on a 900-pixel thumbnail and a 6000-pixel original.

Two shapes of job are covered. A single mark sits at one of nine anchor points — the four corners, the four edge midpoints, or the centre — inset by a margin given as a share of the width, and it can be turned up to ninety degrees either way. Rotation pivots about that anchor rather than about the mark's own middle, so a steep angle in a corner can swing part of the mark past the edge. Ticking the tile box ignores the position and margin altogether and repeats the mark over the whole frame on a lattice tilted by thirty degrees, or by your own rotation when you set one — the arrangement a crop cannot simply cut away. A logo has to arrive as a data: URL rather than a web address, because nothing here is fetched from the network; Image to Base64 turns a PNG into one.

All of it happens in this tab, so the photo is not uploaded and the tool carries on working with the network switched off. The limitation worth weighing is that this is the opposite of the PDF version of the job: the mark is composited into the pixels and cannot be peeled off again, by you or by anyone else, so keep the untouched original. Redrawing on a canvas also discards whatever was attached to the picture rather than painted into it — EXIF, GPS coordinates and the embedded ICC colour profile are all gone from the copy, and a GIF loses its animation, since only the first frame is decoded and the result is written as PNG because canvas cannot encode GIF.

A visible mark is a deterrent rather than a control. A corner stamp can be cropped away, a pale one painted over, and nothing is hidden in the file for later proof; tiling and a lower opacity raise the effort without closing the door. Two smaller things to expect: choosing JPG as the output flattens any transparency in the source onto white, because JPG has no alpha channel and this tool offers no background colour of its own, and there is no live preview, so the first run is where you learn whether the size and placement were right.

How to use it

  1. Add the images. Several can go in at once, and each is stamped in turn with the same settings and returned as its own download. Whatever the browser can decode is accepted: JPG, PNG, WebP, BMP, SVG and GIF, plus AVIF where the browser handles it.
  2. Choose text or a logo. Text is the default and starts at © OmniKitly; an empty box is refused. For a logo, switch to Logo and paste a data: URL beginning data:image/ — an https:// address will not work, as nothing is fetched from the network.
  3. Size and colour the mark. Text runs from 1 to 30 per cent of the image width in sans-serif, serif or monospace, in any colour, with a dark outline switched on by default so pale type stays readable over a bright sky. A logo is sized from 2 to 100 per cent of the width and keeps its own proportions.
  4. Place it, or tile it. Pick one of the nine positions, set the margin from 0 to 20 per cent of the width, and the rotation anywhere between -90 and 90 degrees. Or tick the tile box: position and margin stop mattering, and the mark repeats over the whole picture at a thirty-degree tilt, or at your rotation when that is not zero.
  5. Set opacity and the output format. Opacity runs from 5 to 100 per cent and starts at 55. Keep the input format, or write JPG, PNG, WebP or AVIF; the quality slider starts at 88 and applies to those three lossy formats, never to PNG.

Processing & privacy

Runs on your device. Your file is processed directly in this browser.

Common use cases

Publishing work you would rather not see reposted

A copyright mark low in a corner travels with the file wherever it is saved or reshared, which a caption on the post does not.

Sending proofs to a client

A tiled logo overlay across the set keeps every frame usable for choosing while making them awkward to publish before the invoice is settled.

Branding a batch of listing or catalogue photos

Drop the whole folder in and each shot takes the same mark at the same relative size, so a camera file and a phone snapshot come out matching.

Marking a screenshot before it circulates

Text across the middle at low opacity says which system and which draft a picture came from once it has been forwarded twice.

Questions

Does the picture leave this tab?
No. Decoding, stamping and encoding all happen on a canvas inside this browser tab, so the picture is not sent anywhere and the tool keeps working with the connection off.
Can the mark be removed afterwards?
Not from the file this produces. The mark is composited into the pixels, so there is no layer to switch off and no undo — which is why the original is worth keeping. Someone else can still crop or paint over a mark, and that is the case tiling makes harder.
Why will my logo not load from a web address?
Because this tool never fetches anything from the network. The logo must be pasted as a data: URL starting with data:image/, and Image to Base64 produces one from any PNG. Anything else is refused with a message saying so.
Does it keep EXIF data and the colour profile?
No. The image is redrawn and encoded fresh, so EXIF, GPS coordinates and the ICC profile do not survive into the watermarked copy. That is convenient when the location was going to be stripped anyway, and a reason to keep the original when the camera data mattered.
Can I watermark several images in one go?
Yes. Every file in the queue is processed in turn with the same settings and comes back separately rather than combined. Because text and logo sizes are shares of the image width, the mark stays proportional across files of different dimensions.
What happens to transparent pixels?
They survive into PNG, WebP and AVIF output untouched. JPG has no alpha channel, so they are filled with white before the mark is drawn, and there is no option here to pick a different fill colour.
Is this the right tool for a PDF?
No. Watermark PDF stamps pages as a separate drawing object and leaves the text underneath selectable, which also means it can be deleted again. This one works on raster images and makes the mark permanent.

Related tools