photoprism/photoprism

PhotoPrism – AI‑powered personal photo manager

What it is – PhotoPrism is an open‑source, self‑hosted web application that lets you browse, organize, and share your photos and videos. It uses on‑device machine‑learning models to automatically label images, recognize faces, and extract metadata, all while keeping your data private.

Core capabilities

• Automatic tagging – Deep‑learning models analyse image content and add descriptive labels (e.g., dog, beach, sunset) and location information.
• Face recognition – Detects and groups faces so you can quickly find pictures of family members or friends.
• Powerful search – Combine filters on labels, date, location, resolution, color, quality, etc., to locate any media in seconds.
• RAW & video support – Handles raw camera files and many video formats without conversion.
• Progressive Web App – Works on desktop, tablet, or phone; can be installed like a native app.
• Sync & edit – WebDAV, PhotoSync, and other compatible apps let you back up mobile devices and edit files directly from your computer.
• Privacy‑first – All AI processing runs locally (or on your own server); no data is sent to third‑party cloud services unless you choose to.

How it works (high‑level)

1. Ingestion – You point PhotoPrism at a folder (or sync via WebDAV/PhotoSync). It scans files, extracts EXIF/XMP metadata, and stores thumbnails.
2. Indexing – An embedded TensorFlow‑based model analyses each image, generating tags and face embeddings.
3. Search index – Tags, locations, timestamps, colors, etc., are stored in a searchable database (SQLite/PostgreSQL).
4. Web UI / PWA – The front‑end (Vue.js) queries the index and presents results in a responsive gallery.

Getting started

• Prerequisites – Docker (or Docker‑Compose) and a web browser. The project provides multi‑arch Docker images for AMD64, ARM64, and Raspberry Pi.
• Typical install – docker run -p 2342:2342 -v /path/to/pictures:/photoprism/storage photoprism/photoprism:latest (see the official docs for the full compose file and environment variables).
• Alternative – Download the tar.gz package for Linux/macOS/Windows and run the bundled binary.

Typical use cases

• Personal photo archive for families who want AI‑assisted organization without giving up privacy.
• Small‑business or hobbyist photographers needing a searchable media library on‑premise.
• Anyone who wants a self‑hosted alternative to cloud services like Google Photos.

Limitations / considerations

• AI models are pre‑trained; they work well for common objects and faces but may miss niche categories.
• Initial indexing can be CPU‑intensive for large libraries; a machine with a decent CPU (or optional GPU support) speeds this up.
• The free Community Edition lacks some premium features (interactive world maps, private support) that are unlocked for paying members.
• No built‑in collaborative editing – changes are local to the server instance.

Where to learn more

• Full documentation: https://docs.photoprism.app/
• Public demo (read‑only): https://try.photoprism.app/
• Community chat & discussions links are listed in the README.

---

PhotoPrism is released under the AGPL‑3.0 license; the documentation is CC BY‑NC‑SA 4.0.