Paperless-ngx

Version: 2.20.15
Required minimum memory: 953MB
Last updated 1 May 2026

Recent Changes

  • Update gotenberg to 8.32.0

  • Full Changelog

  • Reverted SSRF defaults (breaking vs 8.31.0). 8.31.0 blocked private-IP destinations by default, which broke deployments running Gotenberg inside a private network. 8.32.0 restores the 8.30.x permissive defaults. Operators with internet-facing APIs opt into the strict posture via the new flags below.

  • Rejected file:// at /forms/chromium/convert/url. Submitting url=file:///tmp/... used to let an unauthenticated caller enumerate the request working directory and read other in-flight uploads as rendered PDFs. The route now returns HTTP 400 for any file:// URL.

  • Required uploaded file for image / pdf stamp and watermark sources. Twelve callsites accepted stampSource=pdf or watermarkSource=pdf with an expression pointing at any path the Gotenberg process could open, even when no file was uploaded. Handlers now return HTTP 400 unless the caller uploaded a matching file.

  • Scoped file:// sub-resources to the request working directory. Crafted HTML could reference another request's file:///tmp/<reqdir>/.... The CDP request handler now restricts file:// sub-resources to the current request's directory. /convert/url and /screenshot/url reject every file:// sub-resource outright.

  • Hardened Chromium against DNS rebinding. A short-TTL DNS authority could return a public IP at validation and a private IP at connect. A loopback HTTP / CONNECT proxy now sits between Chromium and the network, resolves DNS once, and pins the dial to the resolved IP. Skipped when --chromium-proxy-server or --chromium-host-resolver-rules is set.

  • Filtered LibreOffice outbound fetches through a proxy. Uploaded OOXML, RTF, and ODF files can embed external URLs that LibreOffice's libcurl resolves below every Go-side SSRF filter. LibreOffice now routes every outbound fetch through an in-process forward proxy on the same gotenberg.DecideOutbound path Chromium and webhook delivery use. See the four new flags below.

  • Recovered webhook async panics. High-concurrency webhooks could panic the async goroutine and crash the whole process. The goroutine now snapshots the request context and recovers any future panic through the existing error path.

  • LibreOffice outbound URL filtering. Four flags mirror the Chromium and webhook layout: --libreoffice-allow-list, --libreoffice-deny-list, --libreoffice-deny-private-ips, --libreoffice-deny-public-ips. All default permissive.

  • IP-class filtering on four modules. chromium, webhook, api-download-from, and libreoffice each accept matching deny-private-ips and deny-public-ips flags. All default to false.

  • Charts print as blank rectangles (#1531, #1532, #1534, #1535): chromedp v0.15.0 suspended the BeginFrame-driven callback dispatch loop under emulatedMediaType=print. requestAnimationFrame, ResizeObserver, IntersectionObserver, CSS transitionend, and CSS animationend all stopped firing. Pinning chromedp back to v0.14.2 restores native dispatch.

About

Paperless-ngx is an application that manages your personal documents. With the help of a document scanner (see Scanner recommendations), paperless transforms your wieldy physical document binders into a searchable archive and provides many utilities for finding and managing your documents.

Why This Exists

Paper is a nightmare. Environmental issues aside, there’s no excuse for it in the 21st century. It takes up space, collects dust, doesn’t support any form of a search feature, indexing is tedious, it’s heavy and prone to damage & loss.

I wrote this to make “going paperless” easier. I do not have to worry about finding stuff again. I feed documents right from the post box into the scanner and then shred them. Perhaps you might find it useful too.

Paperless-ngx

Paperless-ngx forked from paperless-ng to continue the great work and distribute responsibility of supporting and advancing the project among a team of people.


Install Paperless-ngx in a few minutes on your server with Cloudron. To install Cloudron first, follow our setup steps.
Mastodon Mastodon