Add a text watermark without uploading your photos
Type your name, your brand or your copyright line, pick where it goes, and stamp it on one photo or five hundred — entirely inside your browser. Nine positions, your color and transparency, an optional tiled overlay. If what you want on the picture is a logo rather than a word, this tool stamps the word: the logo's name, a wordmark, or the mark as text. Nothing is uploaded either way.
Why put text on a photo at all? Three reasons come up more than the others. The picture is a draft you are handing out before it is ready, and a word on it stops the person on the other end from reading it as finished. It is a picture going somewhere you do not let anyone use it without your name, and a visible line is the cheapest way to say so. Or it is your work, and the line is the only thing between the file and the next person's website.
- Nothing is uploaded
- JPEG · PNG · WebP · AVIF
- HEIC from iPhone
- Unlimited & free
On a phone? This tool runs there too — grab the QR code and share links →
Drop your images here
or paste from the clipboard · ⌘V / Ctrl+V works too
JPG · JPEG · PNG · WebP · AVIF · HEIC · GIF · BMP — no file count limit
Watermark settings
Applied to every imageThe sample document is generated by code inside your own browser — no image is loaded from the network and nothing is uploaded, so you can safely try the tiled density on it before touching a real ID copy.
A single corner mark can be cropped away; a diagonal tile pattern cannot. Use the corner for aesthetics, the tile when the point is deterrence — or when you are sending a copy of an ID document (see below).
Auto measures each image: it encodes a test pass at several quality levels, scores the result against the original, and keeps the smallest file that still clears a ~40 dB fidelity floor.
Advanced options
The watermark is drawn after any resize, so the text scales with the delivered image — a 4% mark stays 4% no matter the output size.
Auto applies a halo-free unsharp mask only when an image is scaled below 82%.
WASM codecs are fetched from a public CDN on first use and give slightly smaller files. Every fallback still works offline.
Results
What this watermark tool actually does
Every "add watermark" service that runs in a browser tab wants your photos on their server first. This one has nowhere to send them: the picture is decoded, the text is drawn onto it, and the file is re-encoded inside your own browser tab. The same engine powers the site's image compressor and batch resizer — this page just adds one more step to the same local pipeline.
Add one image or several hundred. The watermark settings apply to every image in the batch, files process in parallel background workers, and Download all hands you a single ZIP. The font size is a percentage of each image's shorter edge, so a mixed batch of portraits and landscapes gets consistently proportioned marks rather than one mark that is huge on thumbnails and invisible on posters.
Position, or tile?
A mark in one of the nine grid positions — corners, edge midpoints, or centre — is the aesthetic choice, and a soft shadow keeps it readable on both bright skies and dark shadows. It is also the cropable choice: anything in a corner can be trimmed off. The tiled overlay repeats the text diagonally across the entire frame, which makes casual cropping useless. Marking proofs or paid previews? Use the tile. Signing your own photography? A corner mark usually looks better.
A brand mark, or a transparent watermark
If you are stamping your business on a set of images, the honest answer is that most people type the brand's own word rather than uploading its logo file: the tool draws text, and a wordmark on a soft background reads exactly like a logo mark at the size a photo shows it. Set the Color picker to the brand colour, drop the Opacity to 40–60% for a transparent watermark you can still read the photograph through, and leave the mark in a corner. Bump the opacity back to near solid when the image is going somewhere public and you want the mark to be the first thing someone notices. Either way the photograph underneath is untouched — only the pixels the text covers change.
Protecting copies of ID documents
Whenever you send a photo of a passport, driver's licence or national ID to a bank, landlord or employer, it can be reused by whoever receives it. The standard defence is a full-frame tiled watermark that states the purpose, and this page has the pattern built in: switch on Tile diagonally, set density to Dense, and use text like "For bank account opening only — 30 Sep 2026". Pick a grey or red colour at roughly half opacity, and a size of 5–6% so every line crosses the document number and photo. With lines packed that tight, cropping out one mark leaves the others intact, and the copy becomes useless for anything but the stated purpose.
Because everything happens in your browser, the document itself never touches a server — which is exactly the kind of file you least want to hand to an unknown uploader.
Watermark plus resize plus a KB budget, in one pass
Because the watermark runs on the same pipeline as everything else, you can combine steps that other tools split across uploads: switch Strategy to Target size and the engine watermarks, then searches encoder quality until the file fits the budget — useful for marketplace listings that cap file size. Under Advanced you can cap the maximum edge, so "1280 px wide, JPEG, under 200 KB, watermarked bottom-right" is a single drag-and-drop. The same combination minus the watermark lives on the compress to 200 KB page.
What the mark does to the file
- Metadata — EXIF, GPS coordinates, embedded thumbnails — is stripped in the same pass. If privacy is the goal rather than attribution, the dedicated remove GPS from photo page does metadata removal alone.
- Output format defaults to keeping whatever you put in; JPEG in, JPEG out. Switch to WebP, AVIF or PNG when the destination needs a specific type. PNG files with photography in them can also be palette-reduced — see Advanced.
- The text is burned in — permanently part of the pixels, like any real watermark. That is the point, and also why you should keep clean originals.
Coming from an iPhone? The HEIC to JPG converter accepts the same files and this page handles HEIC directly too. Everything on this site runs on your device; the upload test shows you how to verify that claim against any tool, including this one.
Frequently asked questions
How do I add a text watermark to my photos?
Drop your images on the page, type the watermark text — a copyright line, your name, a brand — pick a position, size, color and transparency, and save. Everything runs in your browser: the text is drawn onto the picture locally and the watermarked files come straight back as downloads.
Can I watermark a whole batch of photos at once?
Yes. Select any number of files, or drag a multi-selection onto the page. The same watermark applies to every image, they process in parallel background workers, and Download all packages the finished files into a single ZIP. The font scales with each image's shorter edge, so mixed sizes get consistently proportioned marks.
Where can the watermark be placed?
Anywhere on a nine-way grid: the four corners, the middle of each edge, or the centre. A soft shadow behind the text keeps it readable on both bright and dark areas. There is also a tiled mode that repeats the text diagonally across the whole frame, in three densities, which is harder to crop out than a single mark.
How do I watermark a copy of my ID, passport or licence?
Switch on Tile diagonally, set Tile density to Dense, and type the purpose into the text — for example "For rental application only — 30 Sep 2026". A grey or red colour at about half opacity keeps the document readable while a size of 5–6% makes each line cross the document number and photo. The dense diagonal grid means no crop can remove every line, so the copy cannot be quietly reused for something else. And since the page runs entirely in your browser, the document is never uploaded anywhere.
Is a watermark enough to protect my photos?
Honestly: no tool can make theft impossible — a watermark in one corner can be cropped away. Put the mark where a crop would hurt the picture, or use the tiled overlay for high-value images, and keep originals elsewhere. What this page does is make proper marking fast enough that you actually do it, on every image, in one pass.
How do I add my company logo to photos?
This tool draws text rather than stamping an image file, so there is no logo upload — but the result people usually want is the same. Type the brand's name or wordmark into Watermark text, set Color to your brand colour, and keep the size around 4% of the shorter edge. If you need the actual mark — the swoosh, the icon, the logotype as a graphic — it has to be a design tool, because stamping a picture onto a picture would be a second decode and a quality loss this page deliberately avoids. What you can do here is put the name on every photo in the batch in one pass, which is the part that takes time by hand.
Can I make the watermark transparent?
Yes — the Opacity control runs from 10% to 100%, so a watermark at 40% is genuinely transparent: the photograph reads straight through the mark and only the letterforms darken the pixels underneath it. Around 60% is the usual balance for a name or brand that should be visible but not dominate. Transparency is per-image in the file it writes; nothing about it depends on the background the mark sits on. Combine it with the tile option and a low density when you want the mark to be felt across the whole frame rather than read in one corner.
Can I resize or compress the photos while watermarking them?
Yes. The watermark is drawn after any resize and before encoding, so the text scales with the delivered image — a 4% mark is 4% of what you download. You can also switch Strategy to Target size to hit a KB budget, and pick JPEG, WebP, AVIF or PNG as the output format.
Are my photos uploaded anywhere?
No. Your files are only ever read by your own browser — there is no server-side component that accepts image data. The only network requests the page makes are for the site's own assets and, optionally, WASM codec modules from a public CDN. Once the page has loaded, watermarking works offline.
