Photofield

Photofield is an open-source, self-hosted photo gallery designed to render and display large volumes of images simultaneously. It functions as a non-invasive viewer that treats the file system as the source of truth, meaning it does not modify original files. This design allows it to be used as a standalone application or as a complementary tool alongside other photo gallery software.

The software is deployed as a single static binary or via Docker images for flexible installation. It indexes local directories to create collections, utilizing a cache database to manage thumbnails and metadata. Users access the gallery through a web browser, where the interface focuses on high-speed loading and flexible viewing layouts to handle thousands of images at once.

Key features

• Progressive multi-resolution loading from low-res previews to full quality
• Zoomable interface for all views to see finer detail
• Semantic search and face detection via photofield-ai
• Reverse geolocation for approximately 50 thousand places using tinygpkg
• Read-only file system based collections to protect original media
• Fast indexing using godirwalk for high-speed file discovery
• Support for videos and multiple resolutions
• Tagging system to filter and search photos

The backend is written in Go and uses SQLite for thumbnail storage, while the frontend is built with Vue 3 and OpenLayers for tiled image rendering. It integrates with FFmpeg for on-the-fly format conversion and can utilize libjpeg-turbo for efficient decoding. The system extracts EXIF metadata and prominent colors as separate follow-up operations to maintain indexing speed. The architecture is optimized for speed on commodity hardware, though it is not designed for a large number of simultaneous clients because much of the state is kept on the server.

Photofield is intended for users who need to browse massive image libraries quickly without altering their existing folder structures or risking the integrity of their original files.