PhotoPrism

Overview

Photos and media pile up across the company — for documentation, project reports, marketing and technical work. To manage, search and organise them efficiently, I introduced PhotoPrism, an open-source media library. The goal: a privacy-friendly, self-hosted system for central image management with modern features like AI-powered image recognition, face detection and automatic categorisation.

Tasks & implementation

Requirements & solution choice

• Defined the requirements for an internal image database
• Chose PhotoPrism for its feature set, privacy-friendliness and Docker compatibility

System preparation

• Provisioned a new VM
• Created a dedicated media storage path via an SMB share

Installation & configuration

• Deployed PhotoPrism with Docker Compose
• Configured user accounts, storage paths and metadata options
• Added automated backups of the media directory

Import & structuring

• Migrated existing image archives (e.g. from Nextcloud and local directories)
• Combined manual tagging with PhotoPrism's automatic tag detection
• Set up albums by department, project and use case

Integration & access

• Integrated it into the existing intranet dashboard (tool overview)
• Defined user groups, roles and permissions
• Trained the departments on usage and upkeep

Result

A modern, intelligent media-management system for the entire company: better access to project image documentation, automatic recognition and sorting by content, place and time, and a single, audit-proof and searchable archive for all photos — privacy-compliant and internal, with no external cloud.

Platforms & tools

• PhotoPrism — open-source photo management
• Docker & Docker Compose — simple, scalable deployment
• SMB share — central media directory
• Nginx / Caddy — reverse proxy with HTTPS