Remove Vidu AI watermark from videos

ClearPix removes the Vidu AI corner wordmark from your own videos directly in the browser. Auto Detect proposes an editable mask for the thin logo or lettering, and local AI inpainting rebuilds the covered pixels in every frame. No sign-up or video upload is required. The export keeps source resolution, duration and audio; use Rect or Paint for faint edges missed by detection.

Erase the thin Vidu AI bottom-right wordmark with automatic selection and local inpainting. Small letters still need a careful edge check.

Drop your video here, or click to browse

MP4 / MOV / WebM / MKV up to 120s — processed locally, never uploaded

Remove watermark from video

Does it work on any watermark position?

Yes. Paint the mask yourself with the brush, or turn on auto-detect to find the watermark on the current frame. Corner logos, center marks, semi-transparent stamps and moving watermarks all work — the model rebuilds whatever pixels your mask covers, frame by frame, entirely on your device.

Is my video uploaded anywhere?

Never. Frames are decoded and processed locally in your browser — your video file stays on your device the whole time and is never sent anywhere. The only download is the AI model (~63 MB), fetched once on first run and cached afterwards for every visit.

How long can the video be?

On the streaming WebCodecs path, clips can run long — the limit is your device’s memory, not a fixed duration. The WASM/ffmpeg fallback is capped at about 30 seconds or 80 MB. If a file is too long, trim it and process it in parts.

Will quality drop?

The output is re-encoded at your original resolution with H.264, so the resolution and frame rate stay the same. Inpainting rebuilds the covered pixels from the surrounding scene; on simple backgrounds the result is clean, while complex textures can soften slightly.

How it works

1

Open your Vidu AI video

Drop the local export you own or have permission to edit. The file stays on your device throughout processing.

2

Check the thin letter edges

Review Auto Detect against the actual wordmark. Cover full glyphs with Rect or Paint without selecting unnecessary surrounding footage.

3

Remove and inspect at full size

Rebuild the marked area across the clip, then look for faint residual strokes along the letters before downloading.

Thin text can be missed without looking missed

All 12 Vidu AI snapshots in our sample used the bottom-right position, and all 12 received a selection in the later image-selection run. That does not prove every thin stroke was covered.

White or semi-transparent letter edges can blend into bright footage. Inspect the mask against both the logo body and its faint outer pixels rather than checking only the box location.

What to expect

  • • Thin or low-contrast glyph edges can require a small manual mask adjustment.
  • • A larger mask is not automatically better over detailed footage.
  • • ClearPix is not affiliated with Vidu AI.

Your file never leaves your device. All AI processing runs locally in your browser on your own hardware — there is no upload, no queue, and no copy stored on any server. You can disconnect from the internet after the page loads and everything still works.

Frequently asked questions

Does this remove both the Vidu logo and text?

Auto Detect aims to select the bottom-right mark as a whole. Adjust the mask for whichever combination of symbol and lettering appears in your export.

Why can faint letter edges remain?

A mask that follows only the strongest part of a thin glyph may miss semi-transparent outer pixels. Slightly extend the selection and run a second pass if needed.

Will a larger box give a cleaner result?

Not necessarily. It must include the full mark, but too much clean footage increases the reconstructed area and can make patching more visible.

Is the Vidu video uploaded?

No. Models and runtime are downloaded to the browser, while the source video is processed locally on your device.

Can I upscale after removal?

Yes. Clean the wordmark first, then upscale the result so the enhancement step does not sharpen residual watermark edges.

Related free tools