Running PhotoPrism on a Raspberry Pi¶
As a result, Raspberry Pi 3 / 4, Apple Silicon, and other ARM64-based devices can pull from the same repository, enjoy the same functionality, and can follow the regular installation instructions after going through a short list of system requirements.
The easiest way to run PhotoPrism on a Raspberry Pi is with PhotoPrismPi.1 Simply flash the image to an SD card, plug it into the Pi and boot it. After a few minutes, our latest release will be ready to use!
- Your device should have at least 3 GB of physical memory and a 64-bit operating system
- While PhotoPrism has been reported to work on devices with less memory, we take no responsibility for instability or performance problems
- RAW image conversion and TensorFlow are disabled on systems with 1 GB or less memory
- Indexing large photo and video collections significantly benefits from local SSD storage and plenty of memory for caching, especially the conversion of RAW images and the transcoding of videos are very demanding
- If less than 4 GB of swap space is configured or a manual memory/swap limit is set, this can cause unexpected restarts, for example, when the indexer temporarily needs more memory to process large files
- High-resolution panoramic images may require additional swap space and/or physical memory above the recommended minimum
- We recommend disabling kernel security in your
docker-compose.yml, especially if you do
not have experience with the configuration:
photoprism: security_opt: - seccomp:unconfined - apparmor:unconfined
- If you install PhotoPrism on a public server outside your home network, always run it behind a secure HTTPS reverse proxy such as Traefik or Caddy
Architecture Specific Notes¶
Modern ARM64-based Devices¶
Running 64-bit Docker images under Raspbian Linux requires a minimum of technical experience to perform the necessary configuration changes. This is because it is a 32-bit operating system with merely a 64-bit kernel to ensure compatibility with legacy software. If you don't need compatibility with 32-bit apps, we recommend choosing a standard 64-bit Linux distribution instead as it will save you time and requires less experience:
- Raspberry Pi Debian
- Ubuntu for Raspberry Pi
- UbuntuDockerPi is a 64-bit Ubuntu Server with Docker pre-configured
Other distributions that target the same use case as Raspbian, such as CoreELEC, will have similar issues and should therefore also be avoided to run modern server applications.
Raspberry Pi OS¶
To ensure compatibility with 64-bit Docker images, your Raspberry Pi 3 / 4 must boot with
arm_64bit=1 flag in its config.txt file.
An "exec format" error will occur otherwise.
Try explicitly pulling the ARM64 version if you've booted your device with the
and you see the "no matching manifest" error on Raspberry Pi OS (Raspbian):
docker pull --platform=arm64 photoprism/photoprism:latest
It may also help to set the
DOCKER_DEFAULT_PLATFORM environment variable to
In case you see Docker errors related to "cgroups", try adding the following parameters to
/boot/cmdline.txt (file location depends on the OS in use):
cgroup_enable=cpuset cgroup_enable=memory cgroup_memory=1
Older ARMv7-based Devices¶
You may use the following 32-bit Docker images to run PhotoPrism and MariaDB on ARMv7-based devices (always use our ARM64 image if possible):
If your device meets the requirements, mostly the same installation instructions as for regular Linux servers apply. Please pay close attention to changed directory and environment variable names.
Darktable is not included in the ARMv7 image because it is not 32-bit compatible. Always choose the regular 64-bit version if your device supports it.
Is a Raspberry Pi fast enough?¶
This largely depends on your expectations and the number of files you have. Most users report that PhotoPrism runs smoothly on their Raspberry Pi 4. However, initial indexing typically takes much longer than on standard desktop computers.
Also keep in mind that the hardware has limited video transcoding capabilities, so the conversion of video file formats is not well-supported and software transcoding is generally slow.
Open a terminal and change to the folder where the
docker-compose.yml file is located.2
Now run the following commands to download the most recent image from Docker Hub and
restart your instance in the background:
docker compose pull docker compose stop docker compose up -d
Pulling a new version can take several minutes, depending on your internet connection speed.
Even when you use an image with the
:latest tag, Docker does not automatically download new images for you. You can either manually upgrade as shown above, or set up a service like Watchtower to get automatic updates.
Our config examples are generally based on the latest stable release to take advantage of performance enhancements. This does not mean older versions are no longer supported and you have to upgrade immediately.
If MariaDB fails to start after upgrading from an earlier version (or migrating from MySQL), the internal management schema may be outdated. See Troubleshooting MariaDB Problems for instructions on how to fix this.
If your device runs out of memory, the index is frequently locked, or other system resources are running low:
- Try reducing the number of workers by setting
PHOTOPRISM_WORKERSto a reasonably small value in
docker-compose.yml, depending on the performance of your device
- Make sure your device has at least 4 GB of swap space so that indexing doesn't cause restarts when memory usage spikes; RAW image conversion and video transcoding are especially demanding
- If you are using SQLite, switch to MariaDB, which is better optimized for high concurrency
- As a last measure, you can disable the use of TensorFlow for image classification and facial recognition
Other issues? Our troubleshooting checklists help you quickly diagnose and solve them.
The default Docker Compose config filename is
docker-compose.yml. For simplicity, it doesn't need to be specified when running
docker-composein the same directory. Config files for other apps or instances should be placed in separate folders. ↩