Original media and sidecar files are scanned for Exif and XMP data, as well as proprietary metadata, including Google Photos JSON. For this, PhotoPrism has a built-in Exif parser, a simple XMP reader and can also use Exiftool to extract metadata in various formats such as Exif, XMP and IPTC:
The combined information is then normalized, merged, and enriched with additional information.
If you update one of these tags with external tools such as Exiftool or Digikam, PhotoPrism reads the changes the next time it indexes the file, provided the file's modification date has been updated.
XMP Sidecar Files¶
When a field is populated with data from an XMP sidecar file, that data is the only source for the field. This means that keywords from XMP override other keywords from PhotoPrism, such as those derived from colors or folder names.
PhotoPrism also reads metadata from Google Photos JSON and Apple XMP files:
In addition to reading metadata from your original and sidecar files, PhotoPrism enriches the metadata of your photos with additional information:
- dates or keywords from folder or filenames
- keywords derived from image classification, color detection and facial recognition
- GPS information from location estimates
- keywords derived from location details
We want you to be able to access your metadata independently of PhotoPrism and its database. That's why the indexer additionally creates human-readable YAML sidecar files that you can open with a text editor or other tools if needed.
PhotoPrism does not yet offer the ability to write changed metadata back to the original files to avoid eventual data loss and conflicts with third-party apps. See GitHub Discussions.