Watermarking Technology

Verda embeds imperceptible watermarks into media so that content can be traced back to its origin — even when metadata has been stripped. Our watermarking technology is designed to survive the transformations that content undergoes as it moves across platforms and devices.

Content watermarking is the process of embedding an imperceptible digital fingerprint into a media file — a photo, video, or audio clip — so that the file can be traced back to its source, even after it has been copied, re-uploaded, screenshotted, or shared across the internet.

Unlike visible watermarks (the translucent logos or text overlays you see on stock photos), an invisible watermark does not alter the appearance or sound of the content in any way a human can perceive. The information is embedded in the signal itself — in pixel values for images, in audio samples for sound, and in video frames for motion content.

Verda uses invisible watermarking to solve a specific problem: proving who published a piece of content first. Every Verda watermark carries a unique identifier tied to a verified creator. When that content surfaces anywhere — on social media, in a news article, on a website — anyone can check the watermark and trace it back to the original verified owner.

Verda's watermarking pipeline processes media in four steps:

1. Upload. You select a photo, video, or audio file in the Verda app and create a post. The original file is uploaded securely to Verda's cloud.
2. Embed. Verda's watermarking engine generates a unique 40-bit identifier for your post and embeds it into the media. For images, the watermark is spread across frequency-domain coefficients. For audio, it is encoded in spectral components below the threshold of human hearing. For video, each frame is individually watermarked.
3. Publish. The watermarked file replaces the original. You can post it to connected social platforms (Instagram, X, LinkedIn, YouTube) directly from Verda, or download it and share it anywhere.
4. Verify. Every watermarked post gets a unique public verification link (e.g., verda.ai/abcdef). Anyone who finds the content later can open that link — or upload the media to Verda's verification tool — to see who the verified creator is, when the content was posted, and where it was originally published.

Visible watermarks — the translucent logos, text overlays, or grid patterns commonly used on stock photography — serve as a deterrent against unauthorized use, but they degrade the viewing experience and are easily removed with modern AI inpainting tools.

Invisible watermarks take a fundamentally different approach. Because the embedded signal is imperceptible, it does not affect how the content looks or sounds. The content can be shared, published, and enjoyed at full quality. At the same time, the watermark persists through the transformations content undergoes online — compression, resizing, format conversion, and re-encoding by social media platforms.

Verda uses invisible watermarking exclusively, because the goal is not to prevent sharing — it is to ensure that whenever content is shared, its origin can be verified.

• Image — PNG, JPEG, AVIF
• Video — MP4, WebM, MOV
• Audio — MP3, WAV, AAC, OGG, FLAC

Our watermarks are designed to persist through common media transformations, including:

• Compression and format conversion
• Resizing and cropping
• Transcoding
• Social media re-encoding

Verda supports C2PA soft binding for manifest recovery. When a platform strips C2PA metadata from content, the embedded watermark enables recovery of the original provenance information. Our algorithms are registered in the C2PA Soft Binding Algorithm List.

Most digital files carry metadata — EXIF data in photos (camera model, GPS, timestamp), ID3 tags in audio, and container metadata in video. Some standards like C2PA go further by attaching cryptographic provenance manifests to files.

The problem: metadata is trivially easy to strip. Every major social media platform removes EXIF data on upload for privacy reasons. C2PA manifests are discarded by platforms that don't support them. A simple "Save As" or screenshot loses all metadata. Once the metadata is gone, there is no way to recover it.

Invisible watermarks solve this because the provenance signal is inside the content itself, not attached alongside it. Stripping metadata does not affect the watermark. The content and the proof of origin travel together, inseparably.

A watermark that ties content to an anonymous account does not solve the impersonation problem. That is why Verda requires every creator to complete a one-time identity verification (KYC) before they can watermark content. This ensures that every Verda watermark is backed by a real, verified person — not a throwaway account.

When someone verifies a Verda-watermarked piece of content, they see not just "this was posted by @handle" but "this was posted by a verified individual whose identity has been confirmed." This is what makes Verda's provenance trustworthy in a way that platform-specific attribution is not.

• Content creators who want to prove ownership of photos, videos, and audio clips they publish online — and trace unauthorized reposts back to the source.
• Photographers and videographers who share work on social media and need a way to prove authorship when their work is copied or re-used without credit.
• Musicians and podcasters who distribute audio content and want to detect when their tracks appear on unauthorized platforms.
• Journalists and news organizations who need verifiable provenance for the media they publish, especially in an era of deepfakes and synthetic content.
• Brands and public figures who want to prevent impersonation and ensure that content attributed to them is genuinely theirs.

What is content watermarking?

Content watermarking is the process of embedding an imperceptible digital fingerprint into media files (photos, videos, audio). Verda's watermarks are invisible to the human eye and inaudible, but can be detected by Verda's verification system to trace content back to its original verified creator.

Can anyone see or hear the watermark?

No. Verda's watermarks are imperceptible — they don't change how your content looks or sounds. The watermark is embedded at a level that is invisible to humans but detectable by Verda's verification technology.

How do I verify if content has a Verda watermark?

Every Verda-watermarked piece of content has a unique verification link (e.g., verda.ai/abcdef). Anyone can open this link to see who the verified creator is. You can also upload media to Verda's verification tool to check for watermarks.

Is Verda free?

Verda offers a free tier with a limited number of watermarks per month. Paid plans (Pro and Team) unlock more watermarks, longer videos, higher resolution, and priority processing. Subscriptions are managed through the App Store (iOS) or Google Play (Android).

Why does Verda require identity verification?

Verda's value is verified provenance — knowing that content comes from a real, verified person. Identity verification ensures that every watermark is tied to a legitimate identity, preventing impersonation and false attribution. It's a one-time process that takes a few minutes.

For questions about our watermarking technology, contact us at support@verda.ai.

Download the Verda app on iOS or Android to start watermarking your content. Free tier available — no credit card required.