PaperStackIf you've looked into self-hosting paperless-ngx, you've probably also found a handful of newer alternatives pitching themselves against it, most of them promising better, smarter, more automatic organization than paperless-ngx's built-in classifier gives you. PaperStack makes a different bet: your digital archive doesn't have to be perfect, because the box still has the original.
What paperless-ngx actually asks of you
paperless-ngx is a capable, well-established open-source tool. It has a large community behind it and a real feature set. What it also asks of you is real: the recommended path is a multi-container Docker Compose stack (webserver, database, broker, optional OCR-support containers), and the bare-metal path needs Tesseract OCR with the correct language files installed, ImageMagick, and a Python virtual environment you're expected to be comfortable with. Some guides that specialize in hosting it for other people put realistic setup time, for someone already comfortable with Docker, at half a day to a full day, and describe the ongoing maintenance (image updates, database migrations, backup checks) as a recurring task that can quietly turn into a second job.
Once it's running, the filing itself is still mostly yours to do. paperless-ngx's built-in classifier only learns from documents you've already tagged and moved out of the inbox, and it retrains on a schedule rather than instantly. That's exactly why a small ecosystem of third-party AI add-ons has grown up around it, built specifically to make automatic tagging more accurate: a real gap in the tool, and the one every paperless-ngx alternative is currently racing to close with better automation.
Where PaperStack disagrees
We're not trying to win that race. PaperStack doesn't promise a smarter classifier or a more accurate OCR pipeline. It promises that when the automatic filing gets something wrong, it doesn't matter as much, because the physical document is still sitting in a labeled box where you put it. Scanning is a batch operation: the whole stack goes through at once, PaperStack splits it back into documents and reads what it can, and the result goes into a searchable index. If a page comes out rough (bad handwriting, a faded receipt, an odd layout), you still have the original a few seconds away.
PaperStack makes that trade-off on purpose. It means PaperStack doesn't need you to design a filename format, doesn't need a folder structure, and doesn't need you to correct its tagging before it becomes useful. You put the stack in, and you can find things by date, by tag, or by anything actually printed on the page.
Side by side
- Setup. paperless-ngx: a Docker stack you host, configure, and keep updated yourself (or pay a managed-hosting provider for). PaperStack: nothing to host, it's a web app, ready when you are.
- Filing accuracy. paperless-ngx: better the more you correct it, with third-party AI add-ons to close the gap further. PaperStack: good enough to search, with the physical box as the fallback when it isn't.
- Ongoing maintenance. paperless-ngx: updates, migrations, and backups are on you. PaperStack: not your problem.
- The paper. paperless-ngx guides often assume you'll eventually shred it. PaperStack assumes you'll keep it, boxed and labeled, on purpose.
Who should actually pick which
If you enjoy maintaining a taxonomy, already run other self-hosted services, and want full control over your document pipeline, paperless-ngx is a genuinely good tool for that. This isn't a case against it. If what you actually want is to stop thinking about filing and still be able to find the one document you need, PaperStack is built around a lower bar: capture fast, search what got read correctly, and fall back to the box for everything else.