PhotoPrism 260728-bbde8f452

With this major new release, equirectangular 360° photos and videos can now be explored interactively. To preserve detail when zooming in, the maximum thumbnail and video resolution has been increased to 16K.

A new inline document viewer makes it possible to read multi-page PDFs without downloading them first. Metadata enthusiasts benefit from enhanced XMP support, allowing GPS coordinates and face regions to be imported from both XMP sidecar files and embedded XMP metadata. In addition, camera and lens make and model information can now be edited through the CLI and API. Elsewhere, the Settings interface has been reorganized and now includes a dedicated Accessibility section, where you can, for example, prevent action menus from opening on hover. Album download and sorting options have moved to a new Collections tab.

AI-related improvements include a new service_tier option for more cost-effective API requests, along with various fixes and improvements for Ollama. A special thank you to all our members, contributors, and testers who helped make this release possible! 🌈

What's new?

• Viewer: Added support for 360° panorama photos and videos by @omerdduran
• Viewer: Added an inline viewer for reading multi-page PDF documents by @omerdduran
• AI: Added service_tier setting for OpenAI-compatible service requests
• AI: Updated Ollama default settings and default cloud model
• AI: Improved Ollama caption and label quality by disabling reasoning output
• AI: Improved retry handling for rate-limited (HTTP 429) vision requests
• AI: Improved image reference validation for computer vision requests
• AI: Fixed a concurrency issue in the local TensorFlow classifier
• UX: Added a setting to prevent action menus from opening on hover by @lastzero
• UX: Improved face-marker overlay performance and interaction by @omerdduran
• UX: Improved notification messages to appear in the current interface language
• UX: Fixed clearing the photo selection after saving batch edits
• Auth: Added a configurable OpenID Connect sign-in prompt to force re-authentication
• Auth: Added OpenID Connect RP-initiated logout to end the upstream provider session
• Auth: Fixed Cognito-issued ID token validation by sending a nonce during OIDC sign-in
• Auth: Fixed the OpenID Connect sign-in button not being translated
• Auth: Improved login, session, and OIDC error messages to appear in the current language
• People: Added import of face regions from XMP sidecar files by @omerdduran
• People: Added import of face regions from embedded XMP metadata by @omerdduran
• People: Added a type-ahead cache for faster name suggestions by @lastzero
• Places: Fixed stacking of photos at the same location when zoomed in
• Albums: Fixed download of sidecar files when the option is enabled
• Albums: Fixed single file downloads from a shared folder, moment, calendar, or region
• Albums: Fixed error when creating a share link as a non-admin user
• Folders: Fixed search results not including pictures from subdirectories
• Batch Edit: Fixed the sorting of pictures in searches after editing dates
• Settings: Added a "Collections" tab to configure download settings by @omerdduran
• Settings: Added 16K thumbnail and video size support for 360° media by @lastzero
• Settings: Reorganized interface with a new accessibility section by @omerdduran
• Index: Added a fallback to embedded JPEG previews for unsupported files
• Index: Added native JPEG XL decoding as an alternative to the external djxl tool
• Index: Fixed duplicate creation when identical files are indexed in parallel by @knowald
• Index: Improved indexing to record original file names only for imported files
• Metadata: Added extraction of GPS coordinates from XMP sidecar files by @omerdduran
• Metadata: Added Lens Make and Model updates via CLI and API by @keif888
• Metadata: Added Camera Make and Model updates via CLI and API by @lastzero
• Metadata: Improved support for metadata from XMP sidecar files by @omerdduran
• Metadata: Improved XMP handling to map dc:subject to the Subject field
• PWA: Added glass, mint, neon, and rainbow app icons and full-bleed touch variants
• PWA: Improved the app manifest with maskable icons, language, and screenshots
• PWA: Fixed using the configured app icon as the iOS Home Screen icon
• CLI: Improved role and auth-provider usage descriptions by @lastzero
• CLI: Fixed auth add command when the database runs with NO_ZERO_DATE enabled
• API: Added an X-Count header to label and service search responses by @keif888
• WebDAV: Fixed uploads reporting success even when they failed
• WebDAV: Improved uploads to skip videos and RAW files when raw and video sync is off
• WebDAV: Fixed routine sync-client folder checks being logged as errors
• Storage: Fixed a path lookup error when resolving filesystem locations
• Storage: Added automatic expiry for temporary download archives
• Database: Improved error logging to capture connection issues
• Database: Upgraded config examples from MariaDB 11.8 to 12.3 (LTS)
• Database: Fixed byte truncation to be rune-safe for all text columns
• Docker: Removed per-user skeleton files to reduce the image size by @alexisLefebvre
• Helm: Added support for external database password secrets by @kurczynski
• Security: Added a feature flag to disable app passwords by @lastzero
• Security: Added signed download tokens and improved preview token handling
• Security: Hardened WebDAV syncing, indexing, and file locking
• Security: Upgraded libheif from v1.22.2 to v1.23.1 (CVE-2026-50142)
• Security: Upgraded Go from v1.26.3 to v1.26.5 and ONNX Runtime to v1.26.0
• Translations: Improved French by @jean-louis67, and Hebrew by @avma

We recommend performing a complete rescan of your library after upgrading to benefit from the improvements. Please note that manually marking faces on 360° content is not yet supported in this release.

360° Photos & Videos

https://github.com/user-attachments/assets/6f4b7dc0-98e5-47c0-babb-7cf5f184e81d

Multi-Page PDF Viewer

Translations

Missing user interface translations have been generated with the help of DeepL and Google Translate. Native speakers are welcome to help us improve them where needed. Learn more ›

Installation Packages

The packages attached to this release are intended for experienced users and maintainers of third-party integrations only, as they require manual configuration and do not include tested system dependencies. Since we are unable to provide support for custom installations, we recommend using one of our Docker images to run PhotoPrism on a private server or NAS device. Learn more ›