MP4/MOV Metadata Inspector
Upload an MP4, MOV, or M4V video and inspect container-level provenance first: ftyp brand, box tree, udta/meta/ilst fields, AIGC labels, common platform strings, SEI markers, and C2PA/JUMBF text clues. This page does not load ffmpeg.wasm on first view; use the homepage detector for sampled frame watermark analysis.
Upload a video to inspect container metadata
How to use
- Choose an original platform export; secondary transcoding often strips container fields.
- Start with the summary and AI-related signals, then expand the box tree and ilst fields.
- If the container has no signal, return to the homepage detector for sampled frame watermark checks.
Fields shown
| ftyp | major_brand, minor_version, and compatible_brands. |
|---|---|
| udta/meta/ilst | Readable encoder, software, tool, comment, AIGC, and similar fields. |
| box tree | MP4/MOV box path, offset, and size for expert field provenance. |
Scope & limits
- This dedicated page prioritizes container inspection and avoids loading ffmpeg.wasm on first view.
- No platform field does not mean the video is not AI-generated; transcoding, editing, and uploads often strip provenance clues.
- For WebM/AVI, start with the homepage detector for basic identification.
FAQ
- How is this different from the homepage AI video detector?
- This page only inspects MP4/MOV file structure and readable container fields. The homepage detector can also sample video frames for watermark signals when needed.
- Can it detect Sora, Kling, or Runway?
- If the original video still carries C2PA-lite, MP4 fields, AIGC labels, SEI markers, or platform strings, they will be shown as provenance clues.
- Why not load ffmpeg immediately?
- The PRD requires a container-first path to avoid heavy frame extraction before the user asks for it, reducing memory and Core Web Vitals risk.