Skip to content

Running PhotoPrism on a Raspberry Pi

Big thank you to Guy Sheffer for building this!

Download our docker-compose.yml file (right click and Save Link As... or use wget) to a folder of your choice, change the configuration as needed, and run sudo docker-compose up to start PhotoPrism:

sudo docker-compose up

See Setup Using Docker Compose for details.

Image name on Docker Hub: photoprism/photoprism-arm64

Operating System and Hardware Requirements

You need to boot your Raspberry Pi 3 / 4 with the parameter arm_64bit=1 in config.txt to be able to use this image. A fast SD card and 4 GB of RAM are recommended, in addition you might want to add swap for large photo collections.


If you're running out of memory while indexing, it often helps to limit the number of workers by setting an explicit value for PHOTOPRISM_WORKERS in docker-compose.yml. Make sure the server has swap configured so that indexing doesn't stop when there are memory usage spikes. As a measure of last resort, you can additionally disable image classification using TensorFlow.

Make sure your docker compose configuration contains the following setting:

      - seccomp:unconfined

Alternatively, you can run the image on UbuntuDockerPi. It's a 64bit Ubuntu Server with Docker pre-installed.

See also: