Two kinds of watermark, one mechanism
You can watermark a video with your own logo or with a stock text mark, and this page supports both by turning them into the same thing.
Both are handled as an image overlaid on the picture. A logo you upload is already an image. A text watermark is drawn to a transparent image in the browser before anything else happens, and from that point on the two paths are identical.
That design is not an aesthetic choice, it comes from a real constraint. The video engine running in your browser has no font files bundled with it, so the filter that would normally draw text has nothing to draw with and would silently produce nothing. Rendering the text in the browser, where fonts obviously do exist, and handing the result over as an image is what makes a text watermark work at all here.
Applying a watermark
- Select your video, and optionally a logo image in the same selection.
- If you did not upload a logo, choose one of the four Text watermark options.
- Choose a Position from the four corners or the centre.
- Set the Opacity as a percentage and the Watermark size as a percentage of video width.
- Run it and download the result, an MP4 named after your video.
Everything is proportional
Both the size and the margin are expressed relative to the video rather than in pixels, for the same reason as on the image version of this tool.
A watermark fixed at 200 pixels wide is a third of the frame on a small clip and a barely visible sliver on a 4K one. Setting it as a percentage of the video’s width means a value of 15 gives the same visual weight regardless of the source resolution, and the height follows from the logo’s own aspect ratio so nothing is ever stretched.
The gap from the edge is 3 percent of the frame, which keeps a corner mark clear of the edge on any resolution and clear of the safe area cropping that some platforms apply.
Why this one cannot avoid a re-encode
Several video tools on this site are near instant because they copy the compressed data through and only rewrite the container. Trimming and splitting work that way, and so does changing the volume, because none of them alters the picture.
A watermark does alter the picture, and there is no shortcut. Every frame has to be decoded, composited with the overlay, and encoded again. That is genuine work and it takes proportionate time on a long clip.
The encode uses a quality target rather than a fixed bitrate, which means the result tracks the source’s own complexity instead of being forced to a size, so the visible quality stays close to the original. It is still one generation of loss, which is unavoidable, and it is why watermarking should be the last step in a workflow rather than the first.
What happens to the sound
Nothing, wherever possible. A watermark has no business touching audio, so an existing AAC track is copied straight across with no re-encode at all.
The exception is a source whose audio codec cannot live in the MP4 container the result uses, which covers the Opus in a WebM and the uncompressed audio in some MOV files. Those are converted to AAC at 128 kbps, which is the minimum change needed to produce a valid file. A silent video comes back silent.
Related watermarking tools
For photos rather than video, Add Watermark to Photos offers the same positions plus a tiled mode, which is worth using when the mark is a deterrent rather than a credit. For documents, Watermark PDF stamps every page.
Because watermarking re-encodes, run Crop Video before it rather than after, so the mark is placed against the final framing. Since the result is always an MP4, Compress Video can take it from there if the file is larger than you want. The rest are on the video hub.