Available now · Docker image published

Run PDFree Inside Your
Own Infrastructure

The same PDF tools your team already uses at pdfree.io, served from a server you control — internal network, no external domain dependency, no per-seat license. For healthcare, legal, finance, and HR teams whose policies require it.

# pull and run — no build step needed
docker pull ghcr.io/mahmudovbahrom555-lab/pdfree-selfhosted:latest
docker run -p 8080:8080 ghcr.io/mahmudovbahrom555-lab/pdfree-selfhosted:latest

Open http://localhost:8080/ — that's it. Full setup, environment variables, and security notes in SELF_HOSTING.md.

Why run it yourself

🔒

Stays on your network

Files already never leave the browser on pdfree.io itself — this goes one step further, so even the page load and static assets never depend on an external domain.

📖

Fully inspectable, AGPLv3

No black box. The exact source that's running is public on GitHub — your security team can read every line before it touches a real document.

💸

Free, no seat count

The image itself has no license fee and no per-user gate. Paid support is optional — the price is below — and nothing in the software depends on it.

A real number, not a marketing claim

Measured directly against this exact image, not estimated: idle memory use, then under a burst of 500 concurrent requests including repeated re-reads of the 1.7MB WASM binary qpdf.wasm that PDFree's own Protect/Unlock tools depend on.

~55 MB
idle
~200 MB
peak under a 500-request burst
~100 MB
settles back to, seconds later

There's a structural reason it stays this low regardless of file size or how many people are using it: the actual PDF processing runs in each visitor's own browser, same as the public site — this container just serves pages, it never opens or processes a single PDF itself. That's a real, testable difference from self-hosted PDF tools that process files inside their own server process, where memory grows with file size and OCR jobs. Verify current numbers directly against whatever you're comparing before deciding — it's a different architecture, not a knock on anyone.

Built for organizations that need it inside their own walls

  • Healthcare IT teams already reading our HIPAA-safety page, who need the tool itself on an internal host
  • Legal and HR departments handling documents that policy says shouldn't touch an external domain at all
  • Finance teams under a data-residency or vendor-review requirement stricter than "processes client-side"
  • Any IT team that simply prefers running its own tools rather than depending on a third-party site's uptime

What's free, and what you can pay for

Free, always

  • The Docker image itself
  • Running it in production, at any scale
  • Reading, forking, modifying the source

Support

$49 a month · or $490 a year

  • Email answers within 2 business days
  • Help getting it running on your server, by email
  • Help with every update
  • An email when a release fixes a security issue
  • Bugs you report are looked at first

Cancel any time. The software keeps working either way — nothing is switched off.

You talk directly to the developer who builds PDFree. We agree the details by email, then send an invoice — no checkout, no seat count. Custom branding or integration work is quoted separately.

Questions IT/procurement ask first

Is the self-hosted image really free?

Yes. PDFree is AGPLv3 — the Docker image is free to pull, run, and even modify, forever, for anyone. There's no license key, no seat count, no feature gate. What you can pay for is optional: a support plan ($49 a month) — never the software itself.

What does AGPLv3 mean for us specifically?

You can run it privately inside your own network with no obligation at all. The one real requirement: if you modify the code and let others interact with your modified version over a network, you must offer them the corresponding source. Running it as-is (the common case) carries no source-sharing obligation beyond what's already public on GitHub.

Does this need internet access to run?

No — once the image is running, PDF processing itself needs no external calls at all (same client-side engine as pdfree.io). The only optional network calls are for a feedback relay you can configure yourself or simply leave disabled.

How is this different from just pointing our team at pdfree.io directly?

Functionally identical tools. What changes is where it's served from — some organizations have policies requiring internal tools to be reachable only on an internal network, or want zero dependency on an external domain even for the initial page load, not just for file processing. This is for that requirement.

Does this include SSO or user accounts?

Not yet. The image is anonymous-by-design, same as the hosted site — there's no built-in authentication layer. Put it behind your own reverse proxy or SSO gateway if you need access control.

What happens if we stop paying for support?

Nothing changes in the software. It keeps running exactly as before, with every tool — you just stop getting email answers and update help. There is no license check to switch anything off.

Ask for support

Leave your email and what you need — you'll get a real reply from the developer, no drip campaign.

Prefer to self-serve? Read SELF_HOSTING.md — no form needed, the image is public right now.

Looking for something else? HIPAA-safety details · Embed a PDF SDK in your own product