How to Add a Watermark to a Video for Free (Text or Logo, No Upload)

You can add a watermark to a video for free in a browser tab, with no upload and no forced watermark of our own. Pick text or a logo, set the corner, size, colour and opacity, then export an MP4 that renders on your device.

The render is FFmpeg compiled to WebAssembly, so the clip is read from your disk into the page and written back as a download. Two facts decide which mode you pick. Text watermarks use one built-in Latin font, so English and accented letters work, but Chinese, Arabic and Cyrillic characters do not; upload a logo image for those. And a watermark has to be drawn onto every frame, so the clip is re-encoded, which takes roughly as long as a compression pass. Text size runs from 2 to 20 percent of the video height, logo width from 5 to 50 percent of the frame, and the position options are the four corners plus centre.

The reason this is worth doing on your own machine is the footage. Clips that need a watermark are usually the ones you would not hand to a random converter site: an internal screen recording, a customer demo, a paid course video or a rough cut that has not shipped yet. An upload-based watermark tool asks you to send that file to a server first, which is a strange place to start when the goal is to mark the file as yours.

Video Watermark

Add text or a logo to any clip, with live preview and five positions. The render happens in the browser, so the video is never uploaded.

Open Video Watermark

What "no upload" changes about watermarking

A watermark is a small change to a big file, and the usual way to make it is to send that big file somewhere. The server receives the clip, decodes it, paints the overlay, encodes it again and sends it back. Everything works, but for the duration of the job your footage is sitting on a machine you do not control, and how long it stays there is decided by someone else's retention policy.

Running the same steps in the browser removes that transfer. The clip is read from local storage into the page memory, the overlay is drawn there, and the result is written back to local storage as a download. There is no queue, no size cap imposed by a hosting plan, and no account. The trade is that the work happens on your device, so a long clip is slower than it would be on server hardware with a GPU.

Two ways to watermark a video

The tool has two modes, and they solve different problems. Text is quick and is what most people want for a name, a handle or a copyright line. An image is what you need for a real logo, or for any wording the built-in font cannot render.

Text watermark

Type up to 60 characters and the text is drawn onto every frame with FFmpeg's drawtext filter. You choose the size as a percentage of the video height, the colour, the opacity and an optional outline that keeps the words legible over busy footage. The default, Aihangsoft, is there only to show the effect; replace it with your own wording. One limit is worth stating plainly: the renderer ships a single Latin typeface, so English and accented Latin letters work and scripts such as Chinese, Arabic, Hebrew, Cyrillic and Greek do not appear at all. For those, use the image mode.

Logo watermark

Switch to Image and upload a PNG or JPG. The overlay is scaled to a width you choose and composited on top of the video, so any font, symbol or drawn graphic works, limited only by the image you supply. A PNG with a transparent background is the cleanest choice because only the visible pixels are laid over the clip. Size runs from 5 to 50 percent of the frame width and opacity from 10 to 100 percent.

How to add a watermark to a video

  1. Add your video. Drop the clip into the tool. MP4, WebM, MKV, MOV and AVI are accepted, and the file is read on your device rather than uploaded. The first run downloads the video engine, about 30 MB, which is then cached.
  2. Set up the watermark. Choose Text or Image. For text, type the wording, pick a size, colour and opacity, and choose an outline if the footage behind it is busy. For a logo, upload the image and set the width. Then pick one of the five positions.
  3. Render and download. Watch the preview to check placement, tap Add watermark, and the exported MP4 appears with a download button. The preview is approximate; the render is drawn at full resolution with the same proportions.

Why adding a watermark means re-encoding

Some video jobs can be done without touching the picture data at all. Trimming often uses stream copy, where the tool simply keeps the frames you asked for and discards the rest, which is nearly instant and lossless. Watermarking cannot work that way, because the pixels of every frame genuinely change. The overlay has to be painted into the image, so the video stream is decoded, composited and encoded again from the first frame to the last.

That is the whole reason the export is slower than a trim. The tool encodes once with H.264 at CRF 20, a quality level chosen so the single pass is hard to notice, and it leaves the resolution and frame rate alone. Audio is copied over as AAC at 128 kbps, and the MP4 is written with its index at the front so it plays while it downloads.

Settings that keep a watermark readable

Position and size matter more than colour. A mark that is too small disappears on a phone screen; one that is too large covers the content it is meant to protect.

ControlRangeDefaultWhen to change it
Text size2 to 20 percent of video height6 percentRaise it for a short handle, lower it for a full legal line
ColourAny colourWhitePick a colour that contrasts with the footage behind it
Text opacity10 to 100 percent70 percentLower it so a mark deters reuse without dominating the frame
OutlineNone, Dark, LightNoneAdd one when the background has both bright and dark areas
Logo width5 to 50 percent of frame15 percentMatch it to how much of the brand mark should stay readable
Logo opacity10 to 100 percent85 percentDrop it for a subtle corner mark over a dark scene
Position4 corners plus centreBottom rightMove it away from captions, faces or an existing platform badge

Placement is the one setting people get wrong most often. The bottom right corner is the default because it is rarely used for content, but a caption track or a UI element in a screen recording can sit exactly there. Check the preview against the busiest part of the clip before you render, not just the first frame.

What this tool does not do

Read this before you start, because a couple of these send people looking for a different tool halfway through.

  • It cannot render non-Latin text as a text watermark. Chinese, Arabic, Hebrew, Cyrillic and Greek are not covered by the built-in font. The workaround is a logo image, not a setting.
  • It always re-encodes. There is no stream copy option, because the picture changes on every frame. Expect the export to take time in proportion to the clip length and resolution.
  • The output is always MP4. WebM, MKV, MOV and AVI inputs are accepted, but the file you get back is H.264 in MP4. Use the Video Converter afterwards if you specifically need another container.
  • One watermark at a time. You choose either text or an image, in one position. There is no tiled repeat and no second overlay.
  • No rotation, animation or timing. The mark is static and covers the whole clip. A watermark that fades in after a few seconds is not something this tool produces.
  • It adds a watermark, it does not remove one. Erasing someone else's watermark from a clip is a different job, and not one this tool does.
  • Short clips are the target. The tool is built for footage under 500 MB and a few minutes long. Longer files are memory bound and slow, and the engine has to hold the clip in the page.
  • No bitrate or codec control. Quality is fixed at CRF 20 and audio at 128 kbps, so if the exported file is too large for its destination, run it through the Video Compressor next.

Frequently asked questions

Yes. The tool is FFmpeg compiled to WebAssembly, so it runs inside the browser tab and the video never leaves your device. There is no upload step, no server queue and no account. You drop the clip in, choose text or a logo, and the rendered MP4 is produced locally and saved as a download. That matters for footage you would rather not hand to a converter site: an internal screen recording, a customer demo, unreleased product footage or a private recording. The first run downloads the video engine, which is about 30 MB of code rather than your file, and it is cached afterwards, so later runs start immediately. You can verify the claim yourself by opening developer tools, switching to the network panel and watching while a watermark renders: no request carries the video.
No. Text watermarks are drawn with one built-in Latin font, which covers English and accented Latin letters such as e, a and u with diacritics, but not Chinese, Arabic, Hebrew, Cyrillic, Greek or Devanagari. The reason is technical rather than a product choice: the video engine runs in a sandbox with no system fonts at all, so the font file has to be shipped with the page, and this page ships one Latin typeface. If your wording needs those scripts, use a logo image instead. Set the watermark type to Image, upload a PNG with the text rendered in any font you like, and the overlay is composited as a picture, which is script independent. A transparent PNG gives the cleanest result because the background of the image is not drawn over the video. For a single short word, an SVG exported to PNG at a generous size keeps the edges sharp.
Very little. The clip is encoded once with H.264 at a quality setting of CRF 20, which is a high quality value rather than a storage saver, and the resolution and frame rate are left exactly as they were. There is no second re-encode and no downscaling, so the visible change is usually impossible to spot on normal footage. What does change is the file size, and it moves in either direction. A phone clip that was already compressed can come out slightly larger, because it is re-encoded at a fixed quality level rather than at its original bitrate; a clip recorded at a very high bitrate can come out smaller. The one visible side effect to expect is on a very low bitrate source, where a single encode pass can soften fine grain and noise. If the source is already close to unwatchable, the watermark will not fix it and may slightly amplify the softness.
A watermark has to be drawn onto every frame of the clip, which means the video stream is decoded, painted and encoded again. That is why it takes time, and the cost scales with length and resolution: expect a short phone clip to finish in under a minute, and a long 4K recording to run for several minutes, because the same work is repeated for every frame. The engine also loads on the first run, which adds a one time download of about 30 MB before the render starts. The progress bar and the log line under the result panel show where the job is, and the tab has to stay open while it works because the compute happens in the page. Trimming the clip first is the cheapest way to shorten the wait, since every second you remove is a second the encoder does not have to process.
Yes. Switch the watermark type from Text to Image and upload a PNG or JPG. A PNG with a transparent background gives the cleanest overlay, because only the logo pixels are drawn over the video and the rest of the image stays clear. Logo width is set as a percentage of the frame, from 5 to 50 percent, and opacity from 10 to 100 percent, so a discreet corner mark and a large centred brand are both a couple of clicks apart. Keep the uploaded file reasonably sized, ideally the width you actually want on screen, because a 4000 pixel logo scaled down to a corner is wasted memory and can slow the render without looking any sharper. The output is always an MP4, whichever way you watermark it, and the original file on your device is never modified.

Watermark your video without uploading it

Free, private and unlimited. No account needed.

Open Video Watermark