You have a drawer (maybe a whole filing cabinet) of documents you cannot throw away. Tax returns. The lease from three apartments ago. An insurance policy you might need someday. A medical record from a specialist visit two years back. You know roughly where everything is until you actually need it, and then the search takes forty-five minutes and ends with you opening envelopes.
Paperless-ngx is an open-source document management system built to end that problem. You scan or photograph a document, drop it into a folder, and Paperless-ngx uses OCR (optical character recognition, the software that reads the text in a scanned image so you can search it later) to extract every word. From that point on, you find any document by typing a keyword, a date, or a name into a search bar. No folders to navigate. No guessing which envelope it is in.
The project has 43,740 stars on GitHub and is the direct successor to the original Paperless and Paperless-ng projects, now maintained by a community team. It released v3.0.4 on July 28, 2026. Full documentation lives at docs.paperless-ngx.com.
| Fact | Value |
|---|---|
| What it is | Self-hosted document management with OCR and full-text search |
| Built by | Community-maintained team; successor to the Paperless project |
| License | GNU GPL v3 (an open-source license; free to self-host, source code must stay open) |
| Platforms | Linux, Raspberry Pi, any machine that runs Docker (works on Intel/AMD and ARM processors) |
| Requires | Docker (software that packages the app into a container you can run on any server or computer), 2 GB RAM minimum |
| Install method | One curl command (interactive script) or Docker Compose |
| Verdict | The right tool for anyone serious about going fully paperless on their own terms |
What Paperless-ngx actually is
Paperless-ngx is a web application that runs on a server or home computer you own. You access it in a browser at something like http://homeserver:8000, and from there you manage every document you have ever scanned or received digitally.
The core mechanic is a consume folder. You put files in it (PDFs, scanned images, Word documents, spreadsheets, photos of receipts) and Paperless-ngx processes them automatically. Processing means three things: running OCR to extract the text from scanned images, converting everything to the PDF/A format (an archival standard designed for long-term storage), and indexing all of it so you can search it instantly.
The metadata model is worth understanding because it is what makes the system actually useful day to day. Every document gets:
- Tags (labels you attach to documents to organize them, like "tax-2025" or "insurance") -- you can apply multiple tags per document
- Correspondent (Paperless-ngx's term for the person or company a document is from: your insurance company, your landlord, the IRS)
- Document type -- a category you define, like "invoice" or "contract" or "medical"
- Date -- parsed from the document automatically when possible
- Custom fields -- free-form metadata you define for your own needs
The machine learning layer is what separates Paperless-ngx from a simple file store. Once you have tagged a few dozen documents, the system learns your patterns and starts suggesting the right tags, correspondent, and document type for new documents automatically. After a few weeks of use, most documents require no manual attention at all.
What you get after installing it
The web interface is a clean dashboard. The document list supports full-text search with auto-completion, relevance sorting, and result highlighting. Search "rent deposit 2023" and you will get the right document even if the scanned image quality was mediocre, because OCR has already extracted the text.
Scanning workflow. Paperless-ngx does not handle the physical scanning step. You still need a scanner or a phone camera for that. But it integrates cleanly with any scanner that supports a network share or email output, and there is a well-maintained mobile app ecosystem (Paperless Mobile for Android, Swift Paperless for iOS) that lets you photograph a document with your phone and send it straight to the consume folder from your couch.
Email importing. You can configure Paperless-ngx to connect to one or more email accounts and automatically pull in attachments. If your bank sends PDF statements, your utility company sends invoices, and your doctor emails visit summaries, all of those arrive in your archive without any manual action from you.
Workflow automation. A built-in workflow engine lets you define rules: if a document matches certain criteria (sender, content keywords, document type), automatically apply tags, assign a correspondent, or trigger a post-processing script. This is the feature that handles the tedium once your archive is established.
Sharing and permissions. The multi-user system supports per-document and per-user access controls. You can share a specific document via a public link, which is useful when a landlord, accountant, or insurance agent asks for a copy and you do not want to dig through email attachments.
AI layer. As of v3.x, optional large language model integration adds document chat (ask a question about a specific document in natural language), smart suggestions for metadata, and similarity retrieval. This is disabled by default and requires a separate AI provider API key.
The install experience
The install script is the easiest path. Paste this into a terminal (curl is a command-line tool that downloads files from the internet; this command downloads and runs the Paperless-ngx setup script):
bash -c "$(curl -L https://raw.githubusercontent.com/paperless-ngx/paperless-ngx/main/install-paperless-ngx.sh)"
The script is interactive. It asks you which database you want (SQLite, a simple file-based database good for personal use, or PostgreSQL, a more robust database server for heavier workloads), where to store documents, whether you want to enable OCR languages beyond English, and a few other questions. At the end it writes a Docker Compose file (a configuration file that describes all the pieces of the app and how they connect) and starts the stack.
The Docker stack has three services: the Paperless-ngx webserver (the main application), a PostgreSQL database (where document metadata lives), and a Valkey instance (a cache service that handles background task queuing, so document processing does not block the web interface). All three are managed together. You do not need to configure any of them individually.
First login takes you through a short setup wizard to configure your admin account. After that, the consume folder is watching. Drop a PDF in, wait thirty seconds, and it appears in your dashboard with the OCR text extracted, ready to search.
Hardware requirements are modest. The project runs well on a Raspberry Pi 4 with 4 GB of RAM, which costs about $55. Many users run it on a spare laptop or a small home server alongside other self-hosted services. The only constraint that matters is disk space: a document archive grows indefinitely, so plan accordingly.
The one genuine friction point is the initial Docker setup if you have never used Docker before. The install script handles the basics, but if something goes wrong (a port conflict, a file permission issue on the storage folder) you are into the documentation. The docs at docs.paperless-ngx.com/setup are thorough, but they assume some comfort with the command line.
Where it fits and what to compare it to
Paperless-ngx occupies a specific niche: a person or household with a backlog of physical documents who wants a permanent, searchable digital archive that does not depend on a subscription service staying in business.
Three honest comparisons:
Evernote / Notion -- both support document uploads and some search, but neither does OCR on scanned images automatically, neither is self-hosted, and both charge monthly fees that compound over years. If your archive has any sensitive documents (tax records, legal agreements, medical history), the question of who else can access your cloud storage is real. Paperless-ngx solves that question permanently: the documents never leave your hardware unless you choose to put them there.
DEVONthink is a powerful macOS-only document manager with excellent OCR and AI features. More capable for complex research workflows, better iOS integration, no server required. But it is a paid macOS application, your archive lives on one machine, and the macOS-only requirement eliminates it for anyone on Linux or Windows. Paperless-ngx is cross-platform, free, and browser-based.
TagSpaces (5,200 stars) is an open-source file organizer that works with your existing folder structure and adds tags. Simpler to install as a desktop app, but it does not do OCR, does not index document content for full-text search, and does not process email. It is a file organizer. Paperless-ngx is a document intelligence system.
The practical decision is simple: if you want a private, searchable archive that processes documents automatically and runs forever without a subscription fee, Paperless-ngx is the right tool. If you want a quick drag-and-drop file store you access from anywhere with no server required, Evernote or Notion will feel faster to start with (and then slower and more expensive over time).
Verdict
Install it. The combination of automatic OCR, machine-learned metadata tagging, email importing, and full-text search covers every use case you will encounter managing a household document archive.
The caveats are real but not dealbreakers. You need to run it on a machine that stays on, which means either a dedicated home server (Raspberry Pi is fine) or a cheap VPS (a virtual private server, a rented slice of a remote computer you control). The initial Docker setup has a learning curve if you have never used Docker before. And the security warning in the README deserves to be repeated: run this only on hardware you trust and control, make regular backups, and do not expose the web interface to the public internet without a VPN (a secure tunnel into your home network) or similar protection in front of it.
None of that changes the core value. Once the archive is built, the hours you spend searching for documents drop to seconds. Tax season becomes filing a folder rather than excavating a drawer. The document you need at the doctor's office, the lease clause your landlord disputes, the receipt for the appliance that just broke: all of them are one search away.
Paperless-ngx is a solved problem you can run for free, on hardware you already own, indefinitely.