What Is Photoview?
Photoview is a self-hosted web gallery built for photographers; its scanner reads a folder on the server and turns the directories into albums. The project calls itself a "photo gallery for self-hosted personal servers", and the code is on GitHub under the AGPL-3.0.
Each user is created together with a path on the server and sees only the media below it. The README links a public demo site.
Get It
Photoview is a server you install yourself; the documented route is Docker Compose, and there is no desktop or phone installer.
Option | How to get it |
|---|---|
| Docker Compose | Compose example: copy the .env, compose file and Makefile, edit them, run make all |
| Other platforms | The README lists Arch Linux AUR, NixOS, Unraid, YunoHost, TrueNAS and EmbassyOS, with limited project support |
| Price | Free, AGPL-3.0, self-hosted |
| Website | photoview.github.io |
| Privacy policy | No privacy policy page was found on the website or in the README |
| Source code | github.com/photoview/photoview |
| Public demo | photos.qpqp.dk with the login shown in the README |
Version reviewed: release 2.4.0, published on 23 June 2024, the newest tagged release; GitHub lists no downloadable files for it. The master branch had commits through 29 October 2026, and the latest and 2.4.0 tags on Docker Hub were last pushed that day. A terminal and Docker are required.
Requirements
The README gives no RAM or CPU minimum; it names supported platforms, databases and a Docker version window.
Runtime
- What the sources say:
- Docker Engine released within the last year; Debian 12 or 13, or Ubuntu LTS and newer short-term releases
Database
- What the sources say:
- SQLite, MariaDB or PostgreSQL 17 or 18
Media processing
- What the sources say:
- RAW and video processing can each be switched off with an environment setting
Video acceleration
- What the sources say:
- Optional qsv, vaapi or nvenc; the project verified qsv on an Intel chip only
These are the README's statements; PromptQuorum measured no memory, CPU or indexing time.
How to Get Started
These steps follow the README's Docker guide; PromptQuorum has not run them.
- 1Install Docker
Why it matters: Install Docker and Docker Compose on the server that holds your photos. - 2Copy the example files
Why it matters: Download the files from the compose example folder, rename them to .env, docker-compose.yml and Makefile, and edit them for your paths and database. - 3Make media readable and start
Why it matters: Run make readable so the container can read your media folders, then make all. - 4Create the admin and scan
Why it matters: Open port 8000, finish the setup wizard with a username, password and photo path (default /photos), then use Scan All in Settings.
Key Features
Every item below comes from the project's website or README; none has been tested by PromptQuorum.
- Sharing. Albums and single items can be shared with a public link, optionally password-protected.
- Maps. Photos with GPS coordinates in their EXIF data can be shown on a map; per the example config this needs a free Mapbox access token that you create.
- Access control. Media files are protected by a cookie token, passwords are hashed and the API uses a strict CORS policy, according to the README.
AI Features and Where They Run
Face recognition is the only AI feature the project lists, and it runs inside the Photoview server using the go-face library.
- Detection and grouping. The scanner finds faces in photos and groups photos of the same person, per the README; the code builds on go-face, an open-source Go wrapper for dlib.
- Model files. Three dlib model files for face detection, landmarks and recognition sit in the repository's data/models folder, and their path is configurable.
- Where it runs. The face-detection code calls the local library directly, with no network request in the files we read.
- Switching off. Setting PHOTOVIEW_DISABLE_FACE_RECOGNITION to 1 skips the feature.
Data Flow and Privacy
Photoview is self-hosted, so photos stay on the server you run unless you share them or turn on maps. The README warns that its basic user authentication is not enough to protect private media from malicious actors on the internet, and suggests a firewall, VPN, reverse proxy or similar layers before exposing it.
The map view loads the Mapbox library in the browser with the token you supply; what Mapbox receives was not checked.
📌Note: These are the project's statements, not audit results. PromptQuorum inspected no network traffic and did not review the code beyond the files named in this article.
Who Should Use It
- Photographers with RAW libraries on a home server who want to browse and share folders without moving files.
- Households that keep family photos on a NAS and want one gallery with a separate folder per user.
- Self-hosters who prefer a folder-based gallery over a platform with AI search and labels.
What We Could Not Verify
- Face recognition quality. Accuracy, grouping behaviour and speed were not tested.
- Release gap. The newest tagged release dates from June 2024 and master has moved on; how far it differs from 2.4.0 was not compared, and the README says PostgreSQL 18 is the recommended version for the next release.
- Mapbox data flow. What the map feature sends to Mapbox, and Mapbox's terms for the free token, were not checked.
- Resource use. No RAM or CPU figures are published, and real usage was not measured.
- Phone experience. No mobile app is listed; how the browser interface behaves on phones was not tested.
Competitors and Alternatives
Three photo-library tools in the directory also recognize faces; they differ in how much else they add.
PhotoPrism
- Platforms:
- Linux, macOS, Windows (Docker), web
- Price / license:
- Free + paid / AGPL-3.0
- Key difference:
- Adds AI labels and search; paid memberships exist
Articles about PhotoPrism (1)
- PhotoPrism Review: Self-Hosted AI Photo Library for Your ServerUpdated October 24, 2026
Immich
- Platforms:
- Linux server, iOS, Android, web
- Price / license:
- Free / AGPL-3.0
- Key difference:
- Adds phone backup apps and smart search
Articles about Immich (1)
- Immich Review: Self-Hosted Photo Library with Smart SearchUpdated October 26, 2026
Also mentioned in:
- Facet Review: Local Photo Scoring, Culling and GalleryUpdated October 28, 2026
Facet
- Platforms:
- Linux, macOS, Windows, web
- Price / license:
- Free / MIT
- Key difference:
- Adds photo scoring and culling; some default model weights are non-commercial
Articles about Facet (1)
- Facet Review: Local Photo Scoring, Culling and GalleryUpdated October 28, 2026
Confirm each app's current price, license and platforms on its own page.
Frequently Asked Questions
Does Photoview change my photo files?
The README describes the gallery as providing read-only access to your media, and generated thumbnails go to a separate cache folder set by PHOTOVIEW_MEDIA_CACHE.
Which database should I use?
SQLite needs no extra service but has the lowest performance and no scalability. The README names MariaDB as the default choice and PostgreSQL as typically the fastest.
How do I update Photoview?
Pull the new image and restart the containers. The README says to back up the database before upgrading, and notes that images moved to the photoview/photoview registry on Docker Hub, so older compose files need the new image name.
Where can I get help?
The README points to a Discord server for setup and development questions, and the project does not support Portainer or other layers on top of Docker Compose.
Verdict
Photoview is an established, openly developed gallery for people who keep photos in folders on their own server and want to browse, share and sort them by face without a cloud service. Against that, the last tagged release is from June 2024, setup needs Docker and a database, face recognition is its only AI feature, maps depend on a Mapbox token, and none of its behaviour could be tested here. It suits folder-minded self-hosters who already run Docker; anyone who wants AI search or phone backup should compare the alternatives above.
Sources
- Photoview website — feature overview, checked 30 October 2026.
- Photoview README — features, platforms, databases, setup and security notes, checked 30 October 2026.
- Photoview releases — release 2.4.0, published 23 June 2024.
- Photoview face-detection code — recognition library and environment settings, checked 30 October 2026.
- Photoview on Docker Hub — image tags, checked 30 October 2026.