PhotoPrism

A self-hosted photo management tool. Upload your photos, and PhotoPrism organizes them — it uses built-in TensorFlow to classify images with labels (e.g. "landscape", "animal", "food"), detect faces, and read GPS data from your camera so you can browse photos by location on a map.

Default Login

• Username: admin
• Password: the password you set in the PHOTOPRISM_ADMIN_PASSWORD variable during deployment

What You Can Do After Deployment

1. Log in — visit your domain and sign in with the credentials above
2. Upload photos — click the upload button (top-right) to drag & drop photos and videos from your computer
3. Automatic classification — PhotoPrism scans your uploads using TensorFlow and assigns labels (e.g. "landscape", "cat", "portrait"), groups faces, and extracts GPS locations
4. Search by label or location — use the search bar to filter photos by labels, dates, locations, or camera model
5. Browse the map — click Places in the sidebar to see your geotagged photos on a world map
6. Organize into albums — select photos and add them to albums for easy browsing

AI Features & Limitations

PhotoPrism includes TensorFlow for offline image classification — this runs locally without GPU and provides basic label detection (objects, scenes) and face grouping. It is not comparable to Google Photos' AI search.

For higher-quality AI (natural language captions, smarter labels), you can optionally connect an OpenAI API key in Settings — this provides the best results at ~3 seconds per image. See AI docs for details.

Note: On containers with 1 GB or less memory, TensorFlow is automatically disabled. Ensure your Zeabur plan provides sufficient memory for AI features.

Uploading from Your Phone

PhotoPrism supports WebDAV, so you can sync photos from your phone:

• Android / iOS: use PhotoSync or any WebDAV-compatible app
• WebDAV URL: https://your-domain.com/originals/

License

AGPL-3.0 — GitHub