April 25, 2025
- Vision API allows remote use of other instances and models
- Arrow keys can be used to navigate in the Edit dialog
- Added additional keyboard shortcuts with usage hints in action menus
- User interface language can be changed on the login page
- Find pictures with specific or no terms in Title and Caption
- Added a collapsible sidebar for displaying metadata in fullscreen viewer
- Added a new menu control in the top bar for additional actions in viewer
- Added Set as Album Cover and Archive/Restore actions to the viewer menu
- Transcoding creates fragmented MP4s optimized for streaming
- Added support for uploading multiple pictures as a zip archive
- Upgraded Google TensorFlow from v1.15.2 to v2.18.0
- Reduced frontend bundle size by 54% for faster loading
- Asynchronous updates improve backend responsiveness
- Improved performance and strategy for manual tagging in people identification
- Result views are reset when there are no matches or an error occurs
- Optimized thumbnail size selection for high aspect ratios
- Improved styles, colors, and mountain shading of the default world map
- Fixed incorrect spelling in log messages
- Improved focus management for more reliable keyboard event handling
- Disabled animation when refreshing search results to prevent flickering
- Fixed opening pictures from randomly sorted search results
This release includes a long list of enhancements and new features, many of them contributed by our community. Most notably, performance has been significantly improved in many areas, we have added a collapsible info sidebar to the fullscreen viewer, it is possible to choose a specific picture as album cover, and you can configure computer vision tasks like image classification to use an external service for scalability and customization. A big THANK YOU to everyone who submitted pull requests, improved translations, or helped with testing! 💖
Important Changes
- To take advantage of the performance improvements in TensorFlow 2, users of our Docker images can set the
PHOTOPRISM_INITvariable to"tensorflow". This will automatically detect, download, and install a version of TensorFlow optimized for your CPU. - If you are using one of our manual installation packages, you can find libtensorflow binaries optimized for AMD64 and ARM64 CPUs on our download server. Updated libheif binaries for HEIC/HEIF image support are also available.
- A 32-bit version for Raspberry Pis with an ARMv7 CPU is no longer part of our regular releases, as TensorFlow 2 does not run on them. We have therefore created an ARMv7 architecture branch with an irregular release schedule and welcome contributions.
What's new?
- AI: Upgraded Google TensorFlow from v1.15.2 to v2.18.0 by @raystlin
- AI: Vision API allows remote use of other instances and models
- UX: Reduced frontend bundle size by 54% for faster loading by @AsikNasik
- UX: Asynchronous updates improve backend responsiveness by @tschechniker
- UX: Arrow keys can be used to navigate in the Edit dialog by @tomplast
- UX: Added additional keyboard shortcuts with usage hints in action menus
- UX: Improved focus management for more reliable keyboard event handling
- UX: Disabled animation when refreshing search results to prevent flickering
- UX: User interface language can be changed on the login page
- UX: Fixed incorrect spelling in log messages by @yarikoptic
- People: Improved performance and strategy for manual tagging by @theshadow27
- Search: Find pictures with specific or no terms in Title and Caption
- Search: Result views are reset when there are no matches or an error occurs
- Viewer: Added a collapsible sidebar for displaying metadata by @omerdduran
- Viewer: Added a new menu control in the top bar for additional actions
- Viewer: Added "Set as Album Cover" and Archive/Restore actions to the menu
- Viewer: Fixed opening pictures from randomly sorted search results
- Viewer: Optimized thumbnail size selection for high aspect ratios
- Videos: Transcoding creates fragmented MP4s optimized for streaming
- Videos: Fixed Intel Quick Sync Video (QSV) hardware transcoding by @hicasper
- Videos: Improved transcoding and indexing logs by @Akashic101
- Places: Improved styles, colors, and mountain shading of the default world map
- Upload: Added support for uploading multiple pictures as a zip archive
- Upload: File extensions and total size of Web uploads can be restricted
- Import: Hidden
.keepand.gitkeepfiles are preserved by @zhzy0077 - WebDAV: Added
Depthheader to improve service discovery by @esteve - Config: Passwords and secrets can be read from files by @petertrr
- Setup: Improved inline documentation in compose.yaml examples
- Docker: Upgraded Ubuntu 25.04 (Plucky Puffin) base image includes FFmpeg v7.1.1, Darktable v5.0.1, RawTherapee v5.11, ExifTool v13.10, and libheif v1.19.7
- Security: Go has been updated to v1.24.2, which includes fixes and improvements
- Translations: Updated Arabic, French, and German
Screenshots
New Viewer Sidebar
yoda-foundtain
Viewer Action Menu
viewer-menu
Improved World Map with 3D Buildings
improved-maps
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.
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.