✅ First Draft Completed 28th May - 8pm

PhotoStructure is a really intriguing entry in our list ?why is it intruiging?. It's a self-hosted photo library management tool which purports to be "your new home for all your photos & videos". Much like the rest of our picks, it's still under heavy development and has some rough edges but the project has a solid foundation to build on as it creeps towards v1.0.

https://s3-us-west-2.amazonaws.com/secure.notion-static.com/af9aa209-7c14-474b-8c5b-994d5bf1d604/Screen_Shot_2021-05-28_at_9.51.31_PM.png

The killer features of this app are based around photo organization. If you're looking for a set it and forget it photo manager which will automatically ingest, deduplicate and organize your photos, run on any major OS whilst transparently mounting volumes and let you be the owner of your own library plus the accompanying metadata forever then this app might just be it.

Right now, there's no machine learning for object analysis or face detection, but these crucial features are on the roadmap. Which brings us neatly to how PhotoStructure development is funded and prioritized. There are two tiers of license: lite and plus. Lite is free and always will be whilst the Plus tier ($5/month or $42/year) gets you full access to all the features the project has to offer. We love that as a lapsed Plus subscriber you library isn't taken hostage behind a paywall, and any prior external work (like automatic organization) will not be undone. You retain access and ownership of Plus level benefits forever.

https://s3-us-west-2.amazonaws.com/secure.notion-static.com/b75f2ad6-dedd-4d2a-b8f1-a0a4baa820a3/Untitled.png

Detailed explanations of settings are par for the course with PhotoStructure

In our testing we found the app to be quite performant, chewing through our sample 10,000 images quite happily overnight. Stability was good and set up was easy (we used the provided docker-compose example). The initial import took several hours as images were de-duplicated, ingested and processed. Given the lack of machine learning features, we can only assume this initial import time will increase. You only do this kind of import once though, right?

As of today, PhotoStructure isn't totally open source, however substantial parts of PhotoStructure are already open source. The author of the software, Matthew McEachen, has stated his commitment to open source the codebase in the event of the parent company closing. In fact when we spoke to Matt we felt confident the project was in the best of hands as when asked about telemetry or user tracking he replied proudly "having a phone home would be antithetical to the goals of the project!". Also by design Matthew told us that PhotoStructure itself doesn't get involved in the business of mobile syncing preferring to steer users toward pre-existing solutions allowing more focus on core feature development.

There are a few drawbacks to PhotoStructure right now. No implementation of any machine learning features yet, it's not fully open sourced and it's another piece of subscription software (is this really a downside to support an indie dev?).

Despite these small nitpicks, the future for the project looks really promising. The voting for new features idea is a great way to ensure the users who are paying the bills get what they want. And if Matthew, who for now at least is a one man band, continues on the current trajectory then this project is certainly one to watch over the coming years.