Self-hosted file storage: what each of the usual options is actually for, and why the simple free one keeps not appearing
The complaint comes up constantly - why is there no simple, free, self-hosted place to put files, the way there is for photos or media. Having run most of the candidates for months at a time, I think the complaint is real but slightly misaimed. Here is where each one actually stops.
The full collaboration suites
The big ones do files and also calendars, contacts, chat, documents, an app store. They work. The cost is that you are running a small office platform to get a folder, and the failure modes are correspondingly large: upgrades that need care, a database, a reverse proxy that must be exactly right, and a background job system that will eventually be why something is not syncing. Nothing here is unreasonable, it is just not what "simple" means.
The plain file servers
A share over the network, exported and forgotten. Enormously reliable, and this is what I would recommend to most people asking. It stops at exactly one thing: getting at your files from outside the house on a phone, without a VPN, without handing anything to the internet. That is the actual gap.
The sync engines
Continuous two-way sync between machines. Excellent at what they do, and people reach for them expecting a drive and get a mirror instead. If a file is wrong on one device, it is now wrong everywhere, quickly. That is not a bug, it is the definition, and it catches people who wanted a backup.
The single-binary web file managers
The closest thing to the request. One process, a folder, a browser. They are genuinely good and the thing they lack is not features, it is the boring half: accounts for other people, sharing links that expire, mobile clients that upload in the background, and someone maintaining it in three years.
Why the simple free one does not appear
Because the easy part is showing a list of files, and every hard part is somewhere else: identity, sharing, mobile background upload, conflict resolution, and enough operational polish to be trusted with the only copy of something. The projects that solve those grow into the suites people are complaining about. The projects that stay simple stay simple by not solving them.
So the honest advice: decide first whether you want a drive, a mirror or a backup, because those are three different products and most of the frustration in this category is someone using one to do another's job.
@sync_is_the_hard · 3w ago · 2 replies
The drive-mirror-backup split is the right frame and I would push it one step further: the reason it collapses into a suite is conflict resolution.
Two devices edit the same file offline. Something has to decide. The plain file server does not have this problem because there is one copy and whoever saves last wins, visibly. The moment you add offline access on a phone you have created the problem, and solving it properly needs versioning, which needs storage policy, which needs a UI, which needs accounts. That chain is why the simple thing becomes the big thing.
Every project in this space that stayed small stayed small by refusing offline mobile editing. Look at any of them and check whether the mobile app is really a viewer with an upload button. It usually is, and that is not laziness, that is the only way to stay out of the chain.
Reply
Report
@conflict_files_again · 3w ago
Conflict resolution being the thing that generates all the complexity is the best sentence in this thread. My evidence is a folder with four hundred conflict copies in it, produced by a tool that was working correctly.
Two devices editing offline is not an edge case, it is Tuesday, and every simple design meets it eventually.
Reply
Report