Verify any media. Images. Audio. Video. Two layers deep.
Upload a file, paste a URL, or enter a Sound ID. TrustMark checks two independent watermark layers — pixel-level steganographic encoding for images and AudioSeal neural watermarking for audio — then surfaces the full provenance chain.
Drop a file here or click to browse
Images (PNG, JPEG, WebP) • Videos (MP4, WebM, MOV) • Audio (MP3, WAV, OGG) — up to 200MB
How dual-layer verification works
Upload any media
Images, audio files, or video — downloaded from a dashboard, screenshot, from social media, or fetched from a URL. Or enter a Sound ID directly.
Two-layer scan
Layer 1 checks pixel-level steganographic data in images. Layer 2 runs AudioSeal neural detection on audio. Both layers are checked simultaneously where applicable.
Full provenance chain
The extracted payload reveals: talent identity, authorizing studio, license status, render receipt ID, Sound ID, and cryptographic signature verification.
Layer 1: Pixel
STEGANOGRAPHICThe full TrustMark payload — receipt ID, license token, talent ID, studio, timestamp — is encoded into the least significant bits of RGB channels. Each pixel carries 3 bits. A 1920×1080 image stores over 770KB of hidden data.
Layer 2: AudioSeal
NEURAL WATERMARKMeta's AudioSeal embeds a 16-bit message into the audio waveform using neural networks. The watermark is imperceptible to human hearing (SNR > 25 dB for speech/music) and recoverable at 99.99% confidence. Each asset gets a unique Sound ID.
Perceptual Fallback
HASH MATCHINGIf compression destroys pixel data, TrustMark falls back to perceptual hash matching. A visual fingerprint is compared against all known watermarked assets. This survives resizing, format conversion, and moderate quality loss.
Try it yourself
Use the Watermark Demo to embed a dual-layer watermark into any image or audio file, then upload the watermarked file here to see the full provenance chain extracted. Or enter the Sound ID directly in the Sound ID tab.