YouTube, Vimeo, Twitter, and TikTok all re-encode uploaded MP4 videos. That re-encoding strips most of the original MP4 atoms — the title, description, copyright, and keyword metadata you carefully embedded. So is there any point in writing MP4 metadata before upload? Yes — for downloads, archives, and the platforms that do preserve it.
YouTube re-encodes every uploaded video to its own codec ladder (H.264, VP9, AV1) at multiple resolutions. The streamed copies that viewers see have YouTube's own metadata pipeline — your embedded ©nam title is replaced by the title you set in YouTube Studio. The original file you uploaded may still have your metadata, but viewers downloading via youtube-dl typically get a stripped copy.
Three real reasons. First, for archive copies on your own server / S3 / DAM where the file is downloaded as-is. Second, for distribution channels that respect metadata (Apple Music podcasts, internal media servers, file-sharing). Third, for legal provenance — embedded copyright + author information travels with the file as evidence of ownership.
Use YouTube's own title, description, and tags fields for YouTube SEO — those are what get indexed. Use embedded MP4 metadata for everything off-YouTube. A clean workflow: write MP4 metadata before upload (for archive integrity), then fill in YouTube Studio fields with the same content (for YouTube SEO). One source of truth, two destinations.
TikTok strips heavily — even the visual codec is replaced. Vimeo preserves more for paid tiers (the source file is downloadable with metadata intact). Twitter/X strips and re-compresses. For consistent behavior across platforms, do not depend on any platform preserving MP4 metadata. Use it for your own archives.