What Metadata Is Stored in Videos?
Published July 14, 2026
Video files carry metadata too, but the details work differently from photos because video formats are containers wrapping multiple streams (video, audio, sometimes subtitles) plus their own metadata layer.
Container-level metadata
Most video containers (MP4, MOV, WebM, MKV) support format-level tags such as:
- Title, description, comment fields
- Creation time — when the file or recording was made
- Encoder — the software or library that produced the file
- Location tags — some phones and cameras write GPS coordinates into video containers, just like they do for photos
- Chapters — timestamped markers within longer recordings
Stream-level details
Independent of the container tags, each stream inside the file carries its own technical characteristics:
- Video stream: codec, resolution, frame rate, bitrate, color information
- Audio stream(s): codec, sample rate, channel layout, bitrate
- Handler names and per-stream tags, which some encoders use to note the specific device or app that produced that stream
Our Video Metadata Checker reports both layers — the container-level tags and each stream's technical details — using ffprobe on a temporary, isolated, auto-deleted server file (there's currently no reliable way to do this fully client-side in a browser).
Provenance and C2PA in video
C2PA's specification covers video containers too, so an MP4 or MOV file can carry a signed manifest the same way a photo can. Our C2PA Checker checks for this directly in your browser via WebAssembly, without uploading the file.
Removing video metadata
Because re-encoding video is slow and can degrade quality, our Video Metadata Cleaner uses a stream copy approach: it remuxes the file with ffmpeg's -map_metadata -1 -c copy, which clears container-level tags and chapters while copying the actual video/audio data through completely unchanged — no re-encoding, no quality loss.
This has real limits worth knowing before you rely on it:
- It can't reach into and edit data embedded inside the encoded stream itself
- It can't guarantee removal of every vendor-specific private box some encoders write, only the metadata tags ffmpeg recognizes
- As a side effect, it often — but not always — drops C2PA manifests during the remux, since ffmpeg typically doesn't carry through container boxes it doesn't understand; the tool verifies this per-file rather than assuming it
If preserving exact video/audio quality isn't a concern and you need a guarantee that specific metadata is gone, treat the before/after report the tool shows you as the source of truth for what actually happened — not this general description.