If you've spent any time generating videos in JiMeng (即梦), you've probably hit the same wall I did. The AI outputs are stunning, especially in Camera Mode where you can insert yourself into a scene with just a photo. But that persistent, semi-transparent watermark crawling across the corner ruins the immersion. It’s there for a reason, of course, but when you’re trying to compile a cohesive portfolio or edit a montage, that logo is a dealbreaker.
The manual workaround—crop it out, blur it, or cover it with a sticker—is tedious. It eats up render time and degrades the actual video quality. The smarter move is to extract the original file that the platform never intended you to see directly. Here’s how to batch-process that workflow without losing your mind or your resolution.
Before we get into the "how," it helps to understand the "why." When you use JiMeng's Camera Mode, you upload a static image or a short clip. The AI then animates it, matching your facial features and movements to a generated scene. This is different from standard text-to-video generation.
Because the output is a derivative of your likeness, the platform applies a stricter watermark—usually a moving logo and a user ID string. Unlike a simple corner stamp, this watermark is often baked into the encoding parameters of the video file itself. This means you can't just "crop" it out without losing significant portions of the frame.
The most efficient way to get a clean file is to bypass the rendered preview entirely. OffWatermark works server-side to pull the source video directly from the CDN. This is the file *before* the watermark layer is composited onto the final output. This is why the quality remains at 100%—there is zero re-encoding involved.
There’s no need to install any heavy desktop software or fiddle with developer console codes. The entire process is browser-based, which makes it perfect for handling multiple videos in a row. Here’s the exact flow I use to clean up a week’s worth of content in under ten minutes.
The biggest time sink is usually going back into JiMeng to find the specific videos. To make this a batch process, you need to gather your links *first*.
This is the crucial step. Don't jump back and forth between apps. Collect five, ten, or twenty links in one sitting. This prep work makes the actual extraction feel like a conveyor belt of clean videos.
Once you have your list, open your browser and head to OffWatermark. No login required for the first few attempts, but creating a free account gets you three extractions immediately.
Repeat this for every link you collected. Because the tool extracts the source file, the download speed is usually faster than re-encoding a video locally on your machine.
Here’s where the "advanced" part comes in. Since OffWatermark pulls the original file, you’re getting the highest bitrate available. But you need to ensure your download settings are correct.
While the title focuses on JiMeng Camera Mode, you’re likely pulling content from multiple sources for a single edit. The beauty of using a multi-platform tool is that you don't have to learn a new workflow for each app.
OffWatermark supports the full ByteDance ecosystem and its competitors. If you're downloading a clip from Douyin (抖音) or the international TikTok, the process is identical. You copy the share link, paste it into the website, and extract.
This is a massive time-saver. For example, if you’re creating a reaction video or a compilation, you might have:
Instead of searching for separate tools for each platform, you run them all through the same pipeline. This consistency is what turns a tedious chore into a smooth workflow. The tool handles the platform-specific API quirks on the backend, so you don't have to think about it.
The goal isn't just to remove a logo; it's to reclaim your time. Manually scrubbing watermarks with editing software takes about 5-10 minutes per video. Using a link extractor takes about 30 seconds per video. When you multiply that by a batch of ten videos, you're saving over an hour of mind-numbing work.
Remember, Dreamina is the international version of JiMeng. If you're using the global version of the app, the watermark removal process works exactly the same. The share link structure is slightly different, but the extraction logic is identical.
A quick note on legality: This tool is designed for creators who want to repurpose their own content. If you generate a video with JiMeng, you own the rights to the creation, but the platform applies a watermark. Removing it to use in your own portfolio or personal project is generally acceptable. However, you should not use this to strip watermarks from other creators' work and pass it off as your own.
If you're tired of the blur tool and ready to get back to editing, give the extraction method a shot.
> Disclaimer: OffWatermark is an independent tool and is not affiliated with, endorsed by, or connected to ByteDance, JiMeng (即梦), Dreamina, Kuaishou (快手), or Xiaohongshu (小红书) in any way. All trademarks belong to their respective owners. Users are solely responsible for ensuring their use complies with applicable laws and terms of service. Only remove watermarks from videos you personally created.
← Back to OffWatermark