For years, Seafile was my go-to solution for file hosting. It all started when I got sick of Google’s ecosystem. I wanted a solution that would offer me a neat (enough) user interface. Seafile works pretty well, and it’s quite easy to set up, but it has some downsides, which led me to leave it for alternatives.

My beef with Seafile is that it stores its data in a custom storage format, making me dependent on whatever Seafile offers for interfacing with my files. There’s also the issue that I had Seafile’s libraries get corrupted because of a system reboot (I’m not completely sure what caused the issue, to be honest). So I had some specific requirements: just a regular filesystem and a bunch of ways to integrate with it.

Those requirements led me to copyparty, which does not look great at all. However, it does offer direct access to files without requiring some kind of interpreter. It offers WebDAV for integrating with it remotely (yes, Seafile does this too, but it feels tacked on). What I also like is that it is pretty fast.

Now, for that copyparty frontend, that’s not something I have gotten used to (yet). So I looked around for an alternative frontend. A WebDAV-powered frontend seemed to make the most sense to me. This is where I found Filestash, which can be the frontend for a bunch of backends, including WebDAV, but S3, FTP, Git, and more are supported as well. It has a nice feature where the authentication can be passed through to the backend, keeping the frontend really simple.

The Filestash + copyparty combo might not be the prettiest, but it gives me something Seafile didn’t: independence regarding my files. My files sit on a normal filesystem, accessible through multiple protocols. If copyparty dies tomorrow, my data isn’t trapped. That is worth living with a less polished interface. For self-hosting, that matters more than it should. Once you’ve experienced data corruption and/or felt the weight of vendor dependency, a plain filesystem starts looking pretty beautiful.

References