Tool
Paperless-ngx
A self-hosted document management system that ingests scans and PDFs, runs OCR, tags and full-text indexes them, and keeps an archive you can actually search — Evernote/Google Drive paperwork-shaped, on disks you control.
What it is
Paperless-ngx is a community-maintained DMS: consume documents from a watch folder or upload UI, OCR them (Tesseract and friends), store originals plus a searchable archive, and organize with tags, correspondents, and document types. People usually run it as a Docker Compose stack (web, consumer, Redis, database). It is not a general file sync suite and not a Photos app.
What it replaces
It replaces Evernote-style clipping/filing, a “PDF dump folder,” and many paid scan-to-search apps for households and small offices that will operate the stack. It does not replace Nextcloud as a full Drive/Talk suite, and it does not replace a tested off-site backup of the archive and database.
License
The paperless-ngx/paperless-ngx repository is licensed under the GNU General Public License v3.0 (GPL-3.0). That is strong copyleft on the software: distributing modified binaries or combining it into a proprietary product has GPL obligations. Bundled OCR engines, fonts, and Docker images can carry their own notices — read those separately when you ship a commercial appliance.
Where it falls over
- OCR quality tracks scan quality and language packs — garbage in stays searchable garbage.
- Consumer + OCR jobs want CPU/RAM; a quiet NAS can spike hard during a backlog ingest.
- The archive is precious; upgrades and restores need a plan before the disk fills.
- Exposing Paperless to the open internet without TLS, auth, and backups is a PII incident.
- GPL-3.0 is not “MIT so we can white-label a SaaS DMS without source obligations.”
Who it's for
People drowning in paper/PDF who will run Docker, schedule backups, and honor GPL. Households and small teams with a real disk budget. Wrong tool if you wanted zero-ops hosted Notes, or a full Nextcloud-shaped suite.
Hardware that usually matters
Paperless’s recurring cost is storage and reliability for the archive — not a gaming GPU. Fast NVMe helps ingest and OCR temp files; capacity matters more than peak TFLOPS. Example listing — not an endorsement; stock and price move.
- Samsung 990 PRO 2TB NVMe SSD — archives and OCR working sets eat disk; a fast NVMe keeps consume jobs from feeling stuck.
Some links are Amazon Associates. We may earn a commission if you buy.
Related in this cluster
- Nextcloud — files/sync suite; not a DMS-first OCR archive
- Immich — photos/videos first, not paperwork
- Docker — how most people actually run Paperless-ngx
- Syncthing — peer folder sync without OCR/search
Verdict
Use Paperless-ngx when you want searchable, self-hosted document archives and you will own OCR quality, GPL, upgrades, and backups. Pair it with real off-site copies of the database and media. Stars are not a substitute for a restore test.