Instead of Paperless
Paperless for the home, the simple version.
Paperless-ngx is a mature, large tool, and for many people it is exactly right. nexpaper is smaller and deliberately plainer: one container, one inbox, one vault per person. This page says honestly which one suits whom, sets the two side by side and shows how the move works if you want to switch.
Two good tools for two different households
Both file paper, read the text and find it again. The difference lies in how much comes with them. Paperless-ngx can do a great deal and can be set up very precisely. nexpaper can do less and wants to manage without a manual.
nexpaper suits you if …
- you want to order the paperwork of a household or a small firm and do not want to run a large setup for it.
- the family should join in, including people who never read a manual.
- you like new things to wait in an inbox first, so that you see them before they are filed.
- everybody should have a vault of their own, with shared ones beside them.
- a single container with text recognition built in is enough.
Paperless-ngx suits you better if …
- you need custom fields, a cost centre for instance.
- you want to decide how files are named on the disk and where they lie (storage paths).
- you want processes that do something on certain triggers (workflows), or matching that learns from your examples.
- you want to file office files and emails as documents.
- you separate stacks with barcodes or want to collect several mail accounts with rules of their own.
Good to knownexpaper is a project of its own. It is not part of Paperless-ngx and does not come from its makers. It is young. Version 1.0 appeared in October 2026. Paperless-ngx, by contrast, is the official successor of Paperless and Paperless-ng and is carried by a community. If you want a tool that has grown over the years, choose Paperless-ngx.
Side by side, line by line
The nexpaper column is right for version 1.0.1. For Paperless-ngx, only things appear that its public documentation describes that way. It is worth reading up, because it keeps developing.
| nexpaper | Paperless-ngx | |
|---|---|---|
| How it is built | one container with SQLite and one worker process, no Redis, no Postgres | several parts: a web server, task processing and a Redis-compatible service, plus a database (PostgreSQL, MariaDB or SQLite); for office files and mails optionally Tika and Gotenberg |
| A new document | waits in the inbox with proposals and their origin until somebody files it | appears in the document list straight away, an inbox tag can mark new things |
| Organising | sender, kind, tags, date, amount and four dates, no custom fields | correspondent, document type, tags, date, custom fields and archive serial number |
| Who sees what | a vault per person, shared vaults, three rights, plus sharing single documents | owners and permissions per object, for users and groups |
| Storage on the disk | a fixed, readable scheme by vault, year and sender | configurable storage paths |
| Matching automatically | rules with plain text comparison, “as last time”, date and amount from the text, AI if you want, none of it learns | matching algorithms per object, among them an automatic one that learns from your documents |
| Processes | none | workflows |
| File types | PDF, pictures, e-invoices as XML, no office files, no mails as documents | office files and mails as well, through Tika and Gotenberg |
| Archive version | a plain searchable PDF, with the original unchanged beside it | can create a PDF/A archive version |
| Mailbox | one IMAP mailbox for the whole server | several mail accounts with rules of their own |
| Separating stacks | at blank separator sheets, scanned on both sides too | at barcodes |
| Saved views | none, the archive has filters | saved views and a dashboard |
| Phone | web app with quick upload, a camera that finds edges, sharing on Android, an iPhone shortcut, no app of its own | a web interface; for phones the documentation points to projects by others |
| Languages of text recognition | six in the image, nothing is downloaded later | many more through Tesseract |
| E-invoices and export | ZUGFeRD, Factur-X and XRechnung are read; a ZIP with PDFs, originals and CSV for the tax adviser | not compared here |
| AI | none out of the box; if you like a local model or a service, only as a proposal | not compared here |
Memory use is hard to compare fairly. At idle nexpaper needs roughly 100 MB of memory, according to a measurement on the developer's test server and only approximately; the peak during text recognition has not been measured again. For Paperless-ngx there is no number here, because it depends on the setup.
What nexpaper leaves out on purpose
Every function that is missing is one button less that somebody has to explain. What stands here, nexpaper has not forgotten. It left it out because hardly anyone at home misses it.
Custom fields
Sender, kind, tags, date, amount and four dates cover a household. If you need more, you get it in the notes, but not as a field you can filter by.
Configurable storage paths
Files always lie in the same scheme, in folders by vault, year and sender, and are named by date and kind. That is readable, and you do not have to invent anything.
Workflows
There are rules when a document is read, but no processes with triggers and actions, and no webhooks either.
Learning matching
nexpaper does not guess from examples. It compares what you tell it, and shows you on every field where the value came from.
Barcodes
nexpaper can separate only at blank sheets. Barcodes or archive serial numbers do not exist.
Office files and mails
nexpaper does not accept Word, Excel and emails. The attachments of a mail do become documents, though.
What is different instead
The inbox where everything new waits, the origin on every field, the vault per person and the readable storage are what nexpaper was built for. On top come page tools with versions, e-invoices and the export for the tax adviser.
What the two have in common
Both run on your own server, and both are free software. With nexpaper the text recognition stays on your own server; there is no service on the internet for it. Whoever runs Paperless-ngx at home loses nothing in the move, because it stays untouched.
How to move from Paperless-ngx
nexpaper fetches documents, senders, kinds, tags and the recognised text through the interface of Paperless-ngx. Nothing is read again, the original in Paperless stays untouched, and you can repeat the move as often as you like. It is done by the person who runs nexpaper.
What comes along, and what does not
- Comes along: the original, Paperless's searchable version, the recognised text, title and date, the correspondent as sender, the document type as kind, tags and notes.
- As a note: custom fields together with the archive serial number become one note with all the values on the document.
- Sharing: What was shared with others in Paperless is shared with the same people in nexpaper.
- Does not come along: storage paths and workflows, because nexpaper has none. Office files come only if Paperless made a PDF of them; then that PDF comes, and the history says so.
What is skipped appears at the end in a list with the Paperless number and the reason.
Step by step
- Prepare
You need Paperless-ngx in version 1.14 or newer, and nexpaper must be able to reach its address. Invite beforehand the people who own documents in Paperless, so that each of them has an account. You do that under “Settings”, “Accounts” with “Invite somebody”.
- Get the token
In Paperless you open your profile and copy the API token. Take the token of an account that may see all documents. Otherwise only what that account sees comes along.
- Connect in nexpaper
Open “Settings”, “Import”. Enter the “Address of Paperless” and the “API token”. If Paperless runs in your home network or on the same machine, tick “Paperless runs in the own network”. nexpaper does not trust a self-signed certificate; in your own network you then use http. A click on “Connect and look” shows how many documents, tags, correspondents and document types there are and which version is running.
- Assign the people
For each owner in Paperless you choose the person in nexpaper. The one with the same name is proposed, otherwise the operator. Under “How it is translated” you see once more what becomes what, and how many new kinds arise.
- Choose the target
Under “Where to” you decide whether each owner gets their documents in their own vault (“Each owner into their vault”) or whether everything goes into one shared vault (“Everything into …”). Everything can be moved afterwards.
- Choose the sorting
“As in Paperless” lets documents that were in Paperless's inbox wait in the inbox and files the others right away. “Everything into the inbox” lets everything wait, and you file it yourself. If you are cautious, take the second and look at the stack in peace.
- Click “Take over N documents”
The move runs in the background. You can close the page, stop it and continue it, and it survives a restart of the server. Whoever received documents is told. At the end it says “Done. N of M taken over.”
- Look at what was skipped
The card “Skipped documents” gives the Paperless number and the reason for each, such as a missing file or a kind nexpaper does not accept.
Good to knowIf something is added in Paperless later, you simply run the move again (“Run again”). That fetches only what is new since. What is in nexpaper already is not fetched twice. Paperless can run alongside for as long as you like. nexpaper takes up to 100,000 documents per run.
And afterwards: tidy what Paperless gave you
After some years of Paperless, similar correspondents often stand side by side. In nexpaper you tidy that up with a few clicks.
Merge senders and tags
On the “Organise” page you see all senders, kinds and tags with the number of their documents. You merge similar ones with “Merge” and rename wrong ones. This works per vault.
Rules for everyday life
You turn habits into rules. Change a sender in the inbox and, at the question “Always like this?”, click “Remember as a rule”. From the next letter on, nexpaper does it by itself.
The storage is readable
In nexpaper the documents lie as ordinary PDF files in folders by vault, year and sender. If you move on one day, you simply take them with you. The way back to Paperless-ngx is open too.
What the history says
What arrives in nexpaper during the move is recorded in the document's history. A document that Paperless made into a PDF from an office file carries a note about it there.