Deploy and Host self hosted Paperless-ngx Matching (Open-Source Document Archive with Auto Matching) on Railway

This template runs Paperless-ngx with Postgres and Redis, built around its matching engine. It suits anyone who wants bills and letters sorted on arrival.

About Hosting Paperless-ngx Matching open-source software on Railway (self hosted Paperless-ngx auto matching and tagging template)

Your documents stay yours. Railway handles provisioning, the generated admin password and the public domain.

Railway also takes the server work off your plate: no OS patching or reverse proxy to configure.

Why Deploy Paperless-ngx Matching, the Evernote alternative on Railway (Railway Free Trial)

Evernote is a closed SaaS note app with OCR search, and it has no matching engine that files documents for you. The $5 Railway trial lets you test it first.

Railway is a singular platform to deploy your infrastructure stack. Railway will host your infrastructure so you don't have to deal with configuration, while allowing you to vertically and horizontally scale it.

By deploying Paperless-ngx Matching on Railway, you are one step closer to supporting a complete full-stack application with minimal burden. Host your servers, databases, AI agents, and more on Railway.

Railway vs Other Hosting Providers and VPS for Paperless-ngx Matching self hosting

Matching needs Postgres, Redis and a container that retrains Auto hourly, so the host matters.

Provider	What You Get with Railway	What You Get with the Other Provider
DigitalOcean	Postgres, Redis and paperless-ngx from one template, with a public https domain set up.	A bare droplet where you handle Docker, backups and the firewall yourself.
AWS	Services and volumes configured in one place, so matching works without extra setup.	Many services to wire together before the first document matches.
Hetzner	One dashboard for the whole archive, at a higher monthly figure for a small archive.	Cheap CPU and storage good for heavy OCR, but Compose and updates are on you.

Pick Railway if you want the archive running fast and would rather tune matching rules than maintain a server.

Common Use Cases for hosted Paperless-ngx Matching

Auto-tagging utility bills: A Fuzzy or Any match on your power company's name files each bill to its correspondent. Misses land in the inbox tag, so you know what needs a look.
Routing tax documents to a storage path: An All match on "W-2" and "1099" sends tax forms to one storage path, which sets their location on disk. Tax-time cleanup gets much shorter.
Letting Auto learn from past filing: After a few hundred correctly tagged documents, switch a tag to Auto and the neural network takes over. Anything still carrying the inbox tag stays out of training.
Catching invoice numbers with regex: A Regular expression match such as INV-\d{6} pulls invoices reliably where a plain word would catch half your archive. Test the pattern on real PDFs first.
Cleaning up the inbox tag workflow: Confirm metadata on each document, then remove the inbox tag so it drops out of the review queue. A short daily pass keeps the queue near zero.

Dependencies for Paperless-ngx Matching Docker hosted on Railway

This template runs three services on Railway's private network: the paperless-ngx web app, Postgres, and Redis as the Celery broker.

Deployment Dependencies for Managed Paperless-ngx Matching Service (Document Archive)

The paperless-ngx service runs the official image ghcr.io/paperless-ngx/paperless-ngx:3.3.0 and serves its web interface on public port 8000.

Implementation Details for Paperless-ngx Matching (Using Paperless-ngx official docker image)

PAPERLESS_SECRET_KEY and PAPERLESS_ADMIN_PASSWORD are generated, so read the admin password from Railway variables to sign in as admin.

How does Paperless-ngx Matching compare against other ways to auto-file scanned documents

Paperless-ngx Matching vs Evernote (Evernote Alternative)

Matching Engine: Paperless-ngx assigns tags, correspondents, document types and storage paths on consumption using Any, All, Exact, Regular expression, Fuzzy or Auto rules. Evernote has no equivalent for filing scans.
Document Focus: Paperless-ngx is a document management system for receipts, bills and letters. Evernote is a note app with OCR search.
Cost Model: Paperless-ngx is GPL-3.0 and runs on your own Railway service. Evernote is a closed SaaS subscription.
Where Evernote Wins: Its sync and apps are more polished, and there is no server to run.

Paperless-ngx Matching vs DEVONthink (DEVONthink Alternative)

Platform: DEVONthink is a macOS and iOS knowledge manager. Paperless-ngx runs in a browser on any machine.
Automation: DEVONthink offers classification on your Mac. Paperless-ngx's Auto mode trains on your tagged documents and files new ones.
Licensing: DEVONthink is proprietary. Paperless-ngx is open source and self hostable.
Where DEVONthink Wins: It is a deeper desktop tool with a richer native interface on Mac.

Paperless-ngx Matching vs Docspell (Docspell Alternative)

Stack: Docspell is an AGPL-3.0 Scala project focused on automation. Paperless-ngx is a Python and Django app with a larger community.
Matching Rules: Both automate tagging, but Paperless-ngx lets you set a match string and algorithm per object, then fall back to Auto.
Where Docspell Wins: Its automation focus can suit heavier processing pipelines, and its AGPL license may matter to you.

Paperless-ngx Matching vs Mayan EDMS (Mayan EDMS Alternative)

Weight: Mayan EDMS is a Django enterprise system with workflows and a heavier footprint. Paperless-ngx is lighter for a personal archive.
Setup Effort: Paperless-ngx matching is set per object in the web UI with one match string. Mayan's workflow model asks for more upfront design.
Where Mayan EDMS Wins: Its workflows and approvals suit organizations that need complex document routing.

How to use Paperless-ngx Matching (the OSS document management system)?

Log in at your public domain as admin with the password from the PAPERLESS_ADMIN_PASSWORD variable in Railway.
Open Correspondents, create one for your bank, pick a matching algorithm, and enter a match string like "Bank of America" BofA.
Do the same for tags and document types. Mark one tag as your inbox tag so you can see what still needs review.
Upload a handful of documents and correct their correspondent, tags and type so the classifier has clean examples.
Leave Auto set on the objects you want learned. The classifier retrains about once an hour, so give it time.
Upload new documents and check which ones picked up their metadata. Documents still carrying the inbox tag are excluded from training, so keep correcting them.

Any:                Bank1 Bank2
All:                invoice total
Exact:              Account Statement
Regular expression: INV-\d{6}

How to self host Paperless-ngx Matching on other VPS Services (Paperless-ngx self hosting guide)

This path is for a plain Linux VPS running the official Docker image.

Clone the Repository

Pull the official repository and move into the Compose folder.

git clone https://github.com/paperless-ngx/paperless-ngx.git
cd paperless-ngx/docker/compose

Install Dependencies

You need Docker and the Compose plugin.

docker --version
docker compose version

Configure Environment Variables

Copy the sample env file, then set your public domain and admin login. PAPERLESS_URL must be the https address so CSRF checks pass.

cp docker-compose.env.example docker-compose.env

PAPERLESS_URL=https://your-domain.example.com
PAPERLESS_SECRET_KEY=replace-with-a-long-random-string
PAPERLESS_ADMIN_USER=admin
PAPERLESS_ADMIN_PASSWORD=replace-with-a-strong-password

Start the Paperless-ngx Matching Application

Bring the stack up in the background, then sign in at your-domain as admin.

Official Pricing of Paperless-ngx Matching (Paperless-ngx pricing)

Paperless-ngx is free and open source under GPL-3.0, with no official paid cloud, so the matching features carry no license fee.

Paperless-ngx Matching cloud vs self hosted comparison (Pricing, features, costs, and more)

Paperless-ngx has no paid cloud from the project, so the real choice is a managed third-party host or running the stack yourself.

Monthly cost of self hosting Paperless-ngx Matching on Railway

On this Railway template you pay for the paperless-ngx service, Postgres, Redis and the volume holding your documents.

System Requirements for Hosting Paperless-ngx Matching on a VPS

The official docs recommend Docker Compose and say Paperless-ngx runs even on a Raspberry Pi, though OCR is slow on weak CPUs.

Frequently Asked Questions (FAQs)

Should I use Auto or a fixed algorithm like Any or Exact?

Use a fixed algorithm when you know the exact wording a correspondent prints, such as Exact or Regular expression for an invoice number.

How often does the Auto classifier retrain?

Auto training runs periodically, by default once each hour, rather than right after you edit a document.

Why did nothing match my new document?

Check that the object has a matching algorithm set and a match string that really appears in the document text.

Is matching case sensitive?

Keep your match strings in mind when testing, because wording and capitalization matter most for Exact and Regular expression.

Does matching work on storage paths too?

Yes, storage paths use the same matching algorithms as tags, correspondents and document types.

Will matching touch documents I already uploaded?

No, matching applies when a document is consumed, so existing documents keep their current metadata.